The Effects of Dual-Tasking on Gait Dynamics in Older ...
Transcript of The Effects of Dual-Tasking on Gait Dynamics in Older ...
The Effects of Dual-Tasking on Gait Dynamics in Older Adults with
Cognitive Impairment
Tess C. Hawkins
Bachelor of Exercise & Sport Science, The University of Sydney
Master of Exercise Physiology, The University of Sydney
A thesis submitted to fulfill of the requirements for the degree of Master of Applied Science
(Research)
Faculty of Health Sciences,
Discipline of Exercise and Sports Science,
The University of Sydney
2019
-i-
SUPERVISORS’S STATEMENT
This is to certify that the thesis entitled “The effects of dual-tasking on gait dynamics in
older adults with cognitive impairment” submitted by Tess C. Hawkins in fulfilment of
the requirements for the degree of Masters by Research is in a form ready for examination.
Professor Maria Fiatarone Singh
Discipline of Exercise & Sport Science
Faculty of Health Sciences
The University of Sydney
-ii-
STATEMENT OF ORIGINALITY
I, Tess C. Hawkins, hereby declare that the work contained within this thesis is my own and
has not been submitted to any other university or institution as a part or a whole requirement
for any higher degree. I certify that the intellectual content of this thesis is the product of my
own work and that all the assistance received in preparing this thesis and sources have been
acknowledged.
In addition, ethical approval from the University of Sydney Human Ethics Committee was
granted for the study presented in this thesis. Participants were required to read a participant
information document and informed consent was gained prior to data collection.
Tess C. Hawkins
Discipline of Exercise & Sport Science
Faculty of Health Sciences
The University of Sydney
-iii-
ACKNOWLEDGMENTS
I would like to extend my sincere thanks and appreciation to all my supervisors Professor
Maria Fiatarone Singh, Dr Yorgi Mavros, Dr Trinidad Valenzuela Arteaga and Professor
Jeffrey Hausdorff. This process has been a continuous learning curve, and I feel incredibly
humbled to have had experienced educators and researchers guiding me throughout,
particularly within the complex world of gait dynamics. Knowing that I had your support was
ever encouraging. You’ve collectively taught me to always raise the bar, and I will continue
to do that.
I would also like to thank the diligent researchers that were part of the Study of Mental And
Resistance Training (SMART) study for their commitment. Likewise, I would like to thank
the participants in the SMART study for their involvement and effort, without their
participation I would not have been able to write this thesis.
Thank you to my research group, a team of generous colleagues committed to improving the
health of others and celebrating every occasion in style. Thank you to my friends in K121 for
providing endless motivation and constantly challenging my opinions.
Most importantly, I would like to thank my family for their unwavering support and
unconditional love. Finally, I would like to thank my wife, Francesca, for being there every
step of the way, acting as my moral sounding board and providing endless reassurance. I
cannot express how none of this would have been possible without you.
-iv-
TABLE OF CONTENTS
SUPERVISORS’S STATEMENT .......................................................................................... i
STATEMENT OF ORIGINALITY ........................................................................................ ii
ACKNOWLEDGMENTS ..................................................................................................... iii
LIST OF TABLES ................................................................................................................. vi
LIST OF FIGURES .............................................................................................................. vii
ABBREVIATIONS ............................................................................................................. viii
ABSTRACT ............................................................................................................................ x
CHAPTER 1: INTRODUCTION ........................................................................................... 1
1.1 RATIONALE ................................................................................................................ 1
1.2 BACKGROUND .......................................................................................................... 2
1.3 FINDINGS FROM PREVIOUS STUDIES ................................................................. 5
1.4 OBJECTIVES ............................................................................................................... 7
1.6 REFERENCES ............................................................................................................. 9
CHAPTER 2: SYSTEMATIC REVIEW.............................................................................. 14
2.1 AUTHOR CONTRIBUTION STATEMENT ............................................................ 15
2.2 ABSTRACT ................................................................................................................ 16
2.3 INTRODUCTION ...................................................................................................... 18
2.4 METHODS ................................................................................................................. 19
2.5 RESULTS ................................................................................................................... 27
2.6 DISCUSSION ............................................................................................................. 39
2.7 STRENGTHS ............................................................................................................. 44
-v-
2.8 LIMITATIONS ........................................................................................................... 44
2.9 CONCLUSIONS......................................................................................................... 45
2.10 REFERENCES ......................................................................................................... 47
CHAPTER 3: STUDY .......................................................................................................... 94
3.1 AUTHOR CONTRIBUTION STATEMENT ............................................................ 95
3.2 PREAMBLE ............................................................................................................... 96
3.3 ABSTRACT ................................................................................................................ 97
3.4 INTRODUCTION ...................................................................................................... 99
3.5 METHODS ................................................................................................................. 99
3.6 RESULTS ................................................................................................................. 107
3.7 DISCUSSION ........................................................................................................... 109
3.8 CONCLUSIONS AND IMPLICATIONS ................................................................ 114
3.9 REFERENCES ......................................................................................................... 115
CHAPTER 4: CONCLUSIONS AND IMPLICATIONS .................................................. 129
4.1 CONCLUSIONS....................................................................................................... 129
4.2 IMPLICATIONS AND FUTURE DIRECTIONS ................................................... 131
4.3 REFERENCES ......................................................................................................... 134
BIBLIOGRAPHY ........................................................................................................... 136
-vi-
LIST OF TABLES
Table 2.1 Medline full electronic search strategy example ........................... 61
Table 2.2 Cohort characteristics ..................................................................... 62
Table 2.3 Risk of bias assessment .................................................................. 66
Table 2.4 Dual-task procedure characteristics ................................................ 68
Table 2.5a Gait outcomes for Mild Cognitive Impairment group: single-task
vs. dual-task comparisons ............................................................... 72
Table 2.5b Gait outcomes for dementia (including Alzheimer’s disease)
group: single-task vs. dual-task comparisons ................................. 76
Table 2.6a Gait outcomes for cognitive status comparisons: Control vs. Mild
Cognitive Impairment ..................................................................... 84
Table 2.6b Gait outcomes for cognitive status comparisons: Control vs.
dementia (including Alzheimer’s disease)...................................... 87
Table 2.7 Gait outcome: within study MCI vs. dementia ............................... 90
Table 2.8 Stratification of outcomes for meta-analysis .................................. 92
Table 3.1 Participant characteristics ............................................................. 126
Table 3.2 Factors significantly associated with changes in at least one
measure of gait variability and dynamics during dual-tasking ..... 127
-vii-
LIST OF FIGURES
Figure 2.1 Flow diagram of the systematic review process ............................. 56
Figure 2.2 Forest plot for within study comparison for single-task and dual-
task for all gait dynamic outcomes and all dual-task procedures .. 57
Figure 2.3 Forest plot for within study comparison for cognitively impaired
and cognitively healthy for all gait dynamic outcomes and all
dual-task procedures ....................................................................... 59
Figure 2.4 Forest plot for meta-analysis of within study comparison for MCI
and dementia for all gait dynamic outcomes and all dual-task
conditions ........................................................................................ 60
Figure 3.1 Gait dynamics and cognitive performance under single-task and
dual-task walking conditions ........................................................ 122
Figure 3.2 Relationship between gait dynamics and brain morphology ........ 124
-viii-
ABBREVIATIONS
a-MCI Amnestic Mild Cognitive Impairment
AD Alzheimer’s disease
AKA Above knee amputation
BKA Below knee amputation
CI Confidence interval
CMT Charcot-Marie-Tooth
CV Coefficient of variations
DFA Detrended fluctuation analysis
DSM-IV Diagnostic and Statistical Manual of Mental Disorders (4th edition)
EF Executive function
EPS Extra-pyramidal signs
ES Effect size
FTD Frontotemporal dementia
GDS Geriatric Depression Scale
HADS Hospital Anxiety and Depression Scale
IEF Impaired executive function
MCI Mild Cognitive Impairment
MD Mean difference
MMSE Mini-mental State Exam
MNA Mini Nutritional Assessment
MS Multiple sclerosis
Na-MCI Non-amnestic Mild Cognitive Impairment
-ix-
NINCDS-ADRDA National Institute of Neurological and Communicative Disorders and
Stroke and the Alzheimer's Disease and Related Disorders
Association
SMD Standardized mean difference
SD Standard deviation
TUG Timed Up and Go
-x-
ABSTRACT
Impaired gait dynamics are associated with increased falls risk, and are worse in cognitively
impaired older adults. Dual-tasking is the performance of a second task, cognitive or physical,
while walking. Dual-tasking impairs gait in individuals with deficits in cognitive function,
and may reveal abnormalities in gait dynamics not observed under single-task conditions,
known as ‘dual-task cost’.
The aims of this thesis were to review the literature on dual-task costs on gait dynamics in
adults with cognitive impairment, and to identify clinical characteristics associated with this
cost in older adults with Mild Cognitive Impairment (MCI) or dementia.
First, a systematic review of 25 articles that measured single- and dual-task walking in adults
with MCI or dementia was conducted. Findings suggested that gait is worsened under dual-
task conditions compared to single-task conditions. Furthermore, dual-task cost is higher in
individuals with cognitive impairment compared to cognitively healthy older adults, and
similarly, higher in adults with dementia compared to adults with MCI. Research is lacking
into nonlinear gait dynamics, the relationship to fall risk, and other characteristics which may
be associated dual-task gait dynamics. Next, data from an acute exposure to dual-tasking in
93 adults with MCI were used to explore linear and nonlinear effects of dual tasking on gait
dynamics. Gait dynamics were assessed using stride time variability and detrended
fluctuation analyses fractal scaling exponent. Cognitive, physical and psychosocial function
and brain morphology were assessed to identify any associations with gait dynamics. Gait
dynamics worsened significantly during dual-tasking, while cognitive performance was
preserved. Additionally, a higher dual-task cost of gait dynamics was associated with lower
-xi-
aerobic fitness, poorer balance, reduced psychological well-being, and reductions in brain
thickness and volume in the posterior cingulate and hippocampus respectively.
Dual-task costs are accentuated in the presence of cognitive dysfunction. Observed
associations with physical fitness, psychological well-being and brain volumes suggest that
interventions targeting these modifiable characteristics could potentially improve dual-task
performance, and ultimately falls risk, in adults with cognitive impairment.
-1-
CHAPTER 1: INTRODUCTION
1.1 RATIONALE
One-third of adults over 65 years of age fall each year [1, 2], and that fall risk is doubled
among older adults with cognitive impairment [3]. Approximately 10% of falls are injurious
[4], and as a leading cause of injury-related hospitalisations [5], the overall cost of falls in the
United States of America is approximately US$50 billion [6], and more than $600 million
per year in Australia [7].
The high global prevalence of falls has prompted research into the prevention and
identification of risk factors associated with falls [8]. Potential risk factors for older adults
include, but are not limited to, sarcopenia, dizziness, gait dysfunction, visual disorders,
postural hypotension, balance impairment and cognitive dysfunction [9]. The multifactorial
nature of falls has been well studied in cognitively healthy older adults, however, less is
known about the nature and interrelationship of risk factors in cognitively impaired older
adults [10].
Variations in gait are important because they contribute to mobility and functional
impairment, and predispose individuals to falls [11]. When a secondary task is added to
walking, cortical control may be challenged and walking regulation may become worse,
leading to a further increase in risk in those already predisposed to fall [12]. Additionally,
changes in cognitive functioning contribute to an increased risk of falls, with risk of falls
heightened under dual-task walking conditions [13]. Dual-task walking is the performance of
a second, concurrent task while simultaneously walking. Therefore, people with cognitive
impairment are more likely to have a gait pattern that is variable, especially under dual-task
-2-
conditions, which may increase their risk of falls. Therefore, this thesis specifically focuses
on gait characteristics in relation to walking under single and dual-task conditions and the
relationships between cognitive impairment and falls risk, as well as demographic,
psychological and physiological factors.
1.2 BACKGROUND
Gait dynamics
A healthy gait appears to flow effortlessly and rhythmically and is characterized by an upright
posture and freely swinging legs [14]. The analysis of gait is commonly performed in clinical
settings to assist in the risk assessment of falls and other neuro-motor outcomes [15].
Measurement methods can vary with assessment choice dependent on the purpose of the
assessment, the cost and usability of equipment, and the assessment environment.
Independent of gait assessment, gait characteristics are usually defined as either spatial,
temporal and, further, as linear nonlinear [16]. Examples of spatial aspects include step length
and step width, while temporal aspects include stride time and swing time. Nonlinear
measures include the unique analysis of temporal gait aspects to determine stability using
fluctuation metrics such as the calculation of a fractal scaling exponent using detrended
fluctuation analyses (DFA) of stride time.
One purpose of gait analysis is to better understand the dynamics of an individual’s gait. Gait
dynamics is a broad term used to describe the magnitude of stride-to-stride fluctuations and
their change over time during walking [4]. It encompasses all aspects of gait variability and
nonlinear changes. Gait variability is identified by intra-individual stride-to-stride
fluctuations in walking parameters such as stride time and step length [17]. Gait variability
assumes larger stride-to-stride fluctuations reflect poorer control of gait. Variations can be
observed during different aspects of the gait cycle and impact the dynamics of gait. Gait
-3-
variability outcomes are commonly measured by distributional metrics, including coefficient
of variation and standard deviation [18]. Distributional metrics are used for linear systems
and quantify the magnitude in variation in a set of spatial or temporal outcomes independently
of the order in the distribution [19]. Alternatively, nonlinear measures quantify the degree of
randomness in highly non-stationary physiological data. Stride interval time series data
exhibit long-range, power law correlations in the gait rhythm [20]. The self-similar, or fractal,
correlations highlight the presence of a ‘memory’ in the neurophysiological locomotor
control system where fluctuations are related to variations in the stride interval hundreds of
strides earlier [4]. Specific tools including DFA, approximate entropy and largest Lyapunov
exponent, have been developed for nonlinear systems to determine variation in how a motor
behavior emerges in time [19, 21]. Nonlinear measures aim to eliminate temporal trends,
which avoids the detection of correlations from non-stationary artifacts.
Recognition of the complex and multifactorial nature of gait dynamics has led to many
investigations to understand why such stride-to-stride or step-to-step variations occur and
what clinical implications they have [17]. Linear and nonlinear measures of gait dynamics
use different methods to assess and quantify the changes in gait variability and stability. As
described above, linear measures capture the magnitude of the changes and nonlinear
measures capture changes over time. The collection of both linear and nonlinear types of gait
outcomes are necessary to completely understand gait dynamics and associated factors.
Cognitive impairment
There is evidence that a more variable gait pattern is associated with cognitive decline in
older adults, particularly with the executive function domain [22]. Executive function (EF)
refers to the cognitive skills responsible for the planning and sequencing goal-oriented tasks
-4-
and the execution of complex activities. Executive function is most often operationalized by
assessments of working memory, attentional controls, and response inhibition [23]. Executive
function impairment has been associated with poorer performance in both spatial and
temporal gait variability measures in community-dwelling older adults [24], as well as in
individuals with the diagnosis of Mild Cognitive Impairment (MCI), dementia and
Alzheimer’s disease (AD) [25]. Mild cognitive impairment reflects a transitional state
between normal aging and AD [26], as well as being a prodrome to other forms of dementia,
including vascular dementia, particularly in subcortical microvascular disease [27].
According to Peterson (1999) [26], MCI can be defined as the following: (1) self-reported
memory complaints, preferably verified by an informant, (2) objective memory disorder, (3)
absence of other cognitive disorders and normal function in everyday life, (4) normal general
cognitive function and (5) absence of dementia. Individuals with MCI experience declines in
cognitive function of 6–10% per year compared to 1–2% in cognitively healthy older adults
[28], resulting in a 10-15 times higher risk of developing AD [29].
The sequencing of cognitive and gait impairments is not clear in MCI as most studies to date
have been cross-sectional, and thus it is not known whether executive function decline
precedes or follows changes in gait stability. Executive function contributes to the
performance of normal walking, and works collectively with sensorimotor systems and other
cognitive domains, i.e., attention, to ensure safe and efficient gait [30]. It is possible that
impaired gait dynamics may be an early sign of brain pathology which precedes the
manifestation of overt cognitive difficulties, and it is thus important to comprehend the
relationship of these two domains. Additionally, given the higher risk of falling in individuals
with cognitive impairment [31], it is important to understand in this cohort specifically,
-5-
whether the performance of complex tasks, involving simultaneous targeting of different
aspects of motor control and cognition, increase fall risk, as reviewed below.
Dual-task walking
Lundin-Olsson and colleagues [2] first reported that nursing home residents who were
observed clinically to stop talking when walking were at a significantly higher risk of falls
over the next 6 months, compared to those subjects who were able to walk and talk. Since
then, dual-tasking, including ‘walking while talking’ has been investigated as a potential
marker of mobility, cognitive impairment and fall risk [32]. Dual-tasking, the performance
of two tasks simultaneously, is a clinically relevant condition that attempts to recreate in the
clinic or laboratory situations where community-dwelling older adults are at a heightened risk
of falling [17]. Dual-tasking has been shown to increase gait variability in adults, especially
in individuals with cognitive impairment. For example, temporal gait variability measures are
significantly worsened under dual task conditions in older adults with MCI and AD compared
to age-matched normal controls [33]. Dual tasking impairments are a known motor
characteristic of MCI, and are a potential marker for further cognitive decline [34].
1.3 FINDINGS FROM PREVIOUS STUDIES
Gait dynamics have been shown to remain relatively stable over time in cognitively healthy
individuals [35] but to worsen over time in cognitively impaired individuals [36].
Additionally, gait dynamics [37] and risk of falls are increased in older adults with MCI
compared to cognitively healthy older adults [38], with those who are cognitively impaired
and have fallen having worse gait dynamics compared to those who are cognitively impaired
but have not fallen [39]. Gait changes under dual-task conditions have also been associated
with future fall risk in cognitively healthy community-dwelling older adults [40]. Linear and
-6-
nonlinear measures of gait dynamics may hold the predictive ability to distinguish between
healthy and fall-prone older adults [16].
Stride time variability [41] and stride length variability [42] have been repeatedly shown to
increase under dual-task conditions compared to single-task conditions. Such dual-task
conditions increase gait variability for both healthy adults and adults with MCI, however,
decrements in performance are different across the cognitive spectrum, with larger
decrements observed in those who are cognitively impaired [33, 37, 43]. Notably, under
dual-task conditions cognitively healthy older adults have been reported to prioritize
cognition over motor performance, referred to as the ‘cognitive-first’ approach [44]. This
pattern is also seen in adults with MCI, who when asked to prioritize either their gait or
cognitive performance during a dual-task experiment, increase gait variability just as they do
with dual-tasking with no instructions regarding task prioritisation [43].
Recent reviews of the literature have focused on the relationships between gait dynamics and
individuals with neurological disorders [18], and use of interventions to target dual-task
performance [45-47]. Several studies implementing interventions targeting dual-task
performance have shown improvements in dual-task gait dynamics in healthy individuals [46]
and individuals with neurodegenerative disease [47]. However, there are too few studies to
generalize the effects of the interventions. Muir-Hunter and colleagues [40] noted a lack of
evidence-based recommendations for dual-task testing to evaluate fall risk in clinical practice.
Finally, research has begun to explore gait characteristics associated with falls in dementia,
including AD [48]. For example, step length variability has been associated with recurrent
falls and the use of mobility aids and walking outdoors, and reduced walking frequency
(amount of time spent walking) has been associated with increased falls risk in individuals
-7-
with dementia [48]. There is great interest in preventing falls by studying a cohort at high risk
for both dementia and gait dynamics under dual-task conditions: older adults with MCI.
Understanding the clinical characteristics which are associated with dual-task gait dynamics
in older adults with MCI is critical to the development of preventative strategies that will
hopefully preserve gait despite intrinsic and environmental stressors known to trigger falls.
1.4 OBJECTIVES
To prevent falls and improve gait dynamics in those at risk, an understanding of the changes
which occur in people with cognitive impairment when dual-tasking is necessary to identify
potential targets for the development of better fall reduction strategies. The aim of this thesis
was to advance the knowledge of dual-task gait dynamics in older adults with mild cognitive
impairment and dementia, and to identify clinical and physiological characteristics associated
with these dynamics.
The following objectives were investigated within this thesis:
1. To evaluate the effect of dual-task walking on changes in gait dynamics, termed dual-
task cost, for older adults with cognitive impairment;
2. To explore if the degree of cognitive impairment, dual-task paradigm and/or gait
dynamic measure influences the dual-task cost of gait dynamics in cognitively
impaired older adults;
3. To identify characteristics associated with the dual-task gait dynamics of cognitively
impaired older adults;
4. To determine the extent of changes in gait and the performance of a secondary
cognitive task during dual-task walking in older adults with cognitive impairment;
-8-
5. To explore associations with dual-task performance/costs, specifically strength,
aerobic capacity, functional performance, psychosocial function, and brain
morphology in adults with cognitive impairment.
-9-
1.6 REFERENCES
1. Blake, A., et al., Falls by elderly people at home: prevalence and associated factors.
Age and Ageing, 1988. 17(6): p. 365-372.
2. Lundin-Olsson, L., L. Nyberg, and Y. Gustafson, “Stops walking when talking” as a
predictor of falls in elderly people. The lancet, 1997. 349(9052): p. 617.
3. Shaw, F.E., Falls in cognitive impairment and dementia. Clinics in Geriatric
Medicine, 2002. 18(2): p. 159-173.
4. Hausdorff, J.M., Gait dynamics, fractals and falls: Finding meaning in the stride-to-
stride fluctuations of human walking. Human Movement Science, 2007. 26(4): p. 555-
589.
5. Pointer, S.C., Trends in hospitalised injury, Australia 1999-00 to 2014-15, in Injury
research and statistics series no. 110. 2018, AIHW: Canberra.
6. Florence, C.S., et al., Medical Costs of Fatal and Nonfatal Falls in Older Adults.
Journal of the American Geriatrics Society, 2018. 66(4): p. 693-698.
7. Bradley, C. and J. Harrison, Hospitalisations due to falls in older people, Australia,
2003-2004, in Injury research and statistics series no. 32. 2007, AIHW: Canberra.
8. Stewart Williams, J., et al., Prevalence, risk factors and disability associated with
fall-related injury in older adults in low- and middle-incomecountries: results from
the WHO Study on global AGEing and adult health (SAGE). BMC Medicine, 2015.
13(1): p. 147.
9. Rubenstein, L.Z., Falls in older people: epidemiology, risk factors and strategies for
prevention. Age and Ageing, 2006. 35(S2): p. 37-41.
10. Montero-Odasso, M., et al., Gait and cognition: A complementary approach to
understanding brain function and the risk of falling. Journal of the American
Geriatrics Society, 2012. 60(11): p. 2127-2136.
11. Hausdorff, J.M., D.A. Rios, and H.K. Edelberg, Gait variability and fall risk in
community-living older adults: A 1-year prospective study. Archives of Physical
Medicine and Rehabilitation, 2001. 82(8): p. 1050-1056.
-10-
12. Montero-Odasso, M., et al., Dual-tasking and gait in people with mild cognitive
impairment. The effect of working memory. BMC Geriatrics, 2009. 9(41).
13. Lamoth, C.J., et al., Gait stability and variability measures show effects of impaired
cognition and dual tasking in frail people. Journal of Neuroengineering &
Rehabilitation, 2011. 8: p. 2.
14. Pirker, W. and R. Katzenschlager, Gait disorders in adults and the elderly. Wiener
Klinische Wochenschrift, 2017. 129(3): p. 81-95.
15. Maki, B.E., Gait Changes in Older Adults: Predictors of Falls or Indicators of Fear?
Journal of the American Geriatrics Society, 1997. 45(3): p. 313-320.
16. Hamacher, D., et al., Kinematic measures for assessing gait stability in elderly
individuals: a systematic review. Journal of the Royal Society Interface, 2011. 8(65):
p. 1682-1698.
17. Hausdorff, J.M., Gait variability: methods, modeling and meaning. Journal of
NeuroEngineering and Rehabilitation, 2005. 2(1): p. 19.
18. Moon, Y., et al., Gait variability in people with neurological disorders: A systematic
review and meta-analysis. Human Movement Science, 2016. 47: p. 197-208.
19. Stergiou, N. and L.M. Decker, Human movement variability, nonlinear dynamics, and
pathology: Is there a connection? Human Movement Science, 2011. 30(5): p. 869-
888.
20. Hausdorff, J.M., et al., When human walking becomes random walking: fractal
analysis and modeling of gait rhythm fluctuations. Physica A: Statistical Mechanics
and its Applications, 2001. 302(1): p. 138-147.
21. Hausdorff, J.M., et al., Fractal dynamics of human gait: stability of long-range
correlations in stride interval fluctuations. Journal of Applied Physiology, 1996.
80(5): p. 1448-1457.
22. Byun, S., et al., Gait Variability Can Predict the Risk of Cognitive Decline in
Cognitively Normal Older People. Dementia and Geriatric Cognitive Disorders,
2018: p. 251-261.
-11-
23. Royall, D.R., et al., Executive Control Function. The Journal of Neuropsychiatry and
Clinical Neurosciences, 2002. 14(4): p. 377-405.
24. Martin, K.L., et al., Cognitive Function, Gait, and Gait Variability in Older People:
A Population-Based Study. The Journals of Gerontology: Series A, 2013. 68(6): p.
726-732.
25. Yogev-Seligmann, G., J.M. Hausdorff, and N. Giladi, The role of executive function
and attention in gait. Movement Disorders, 2008. 23(3): p. 329-342.
26. Petersen, R.C., Mild Cognitive Impairment: Clinical Characterization and Outcome.
The Journal of the American Medical Association, 1999. 281(19): p. 303-308.
27. Meyer, J.S., et al., Is Mild Cognitive Impairment Prodromal for Vascular Dementia
Like Alzheimer's Disease? Stroke, 2002. 33(8): p. 1981-1985.
28. Lehrner, J., et al., Annual conversion to Alzheimer disease among patients with
memory complaints attending an outpatient memory clinic: the influence of amnestic
mild cognitive impairment and the predictive value of neuropsychological testing.
Wiener Klinische Wochenschrift, 2005. 117(18): p. 629-635.
29. Petersen, R.C., et al., Current concepts in mild cognitive impairment. Archives of
Neurology, 2001. 58(12): p. 1985-1992.
30. Snijders, A.H., et al., Neurological gait disorders in elderly people: clinical approach
and classification. The lancet neurology, 2007. 6(1): p. 63-74.
31. Muir, S.W., K. Gopaul, and M.M. Montero Odasso, The role of cognitive impairment
in fall risk among older adults: a systematic review and meta-analysis. Age and
Ageing, 2012. 41(3): p. 299-308.
32. Ayers, E.I., et al., Walking while Talking and Falls in Aging. Gerontology, 2014.
60(2): p. 108-113.
33. Muir, S.W., et al., Gait assessment in mild cognitive impairment and Alzheimer's
disease: The effect of dual-task challenges across the cognitive spectrum. Gait and
Posture., 2012.
-12-
34. Montero-Odasso, M., et al., The motor signature of mild cognitive impairment: results
from the gait and brain study. Journals of Gerontology Series A-Biological Sciences
& Medical Sciences, 2014. 69(11): p. 1415-1421.
35. Herssens, N., et al., Do spatiotemporal parameters and gait variability differ across
the lifespan of healthy adults? A systematic review. Gait & Posture, 2018. 64: p. 181-
190.
36. van Iersel, M.B., et al., Systematic review of quantitativeclinical gait analysis in
patients with dementia. Zeitschrift fur Gerontologie und Geriatrie, 2004. 37(1): p. 27-
32.
37. Montero-Odasso, M., S.W. Muir, and M. Speechley, Dual-task complexity affects gait
in people with mild cognitive impairment: the interplay between gait variability, dual
tasking, and risk of falls. Archives of Physical Medicine & Rehabilitation, 2012.
93(2): p. 293-299.
38. Liu-Ambrose, T.Y., et al., Increased risk of falling in older community-dwelling
women with mild cognitive impairment. Physical Therapy, 2008. 88(12): p. 1482-
1491.
39. Allali, G., et al., Falls, Cognitive Impairment, and Gait Performance: Results From
the GOOD Initiative. Journal of the American Medical Directors Association, 2017.
18(4): p. 335-340.
40. Muir-Hunter, S.W. and J.E. Wittwer, Dual-task testing to predict falls in community-
dwelling older adults: a systematic review. Physiotherapy, 2016. 102(1): p. 29-40.
41. Al-Yahya, E., et al., Cognitive motor interference while walking: A systematic review
and meta-analysis. Neuroscience and Biobehavioral Reviews, 2011. 35(3): p. 715-
728.
42. Smith, E., et al., The Influence of a Cognitive Dual Task on the Gait Parameters of
Healthy Older Adults: A Systematic Review and Meta-Analysis. Journal of Aging &
Physical Activity, 2017. 25(4): p. 671-686.
-13-
43. Lee, J. and S. Park, Effects of a priority-based dual task on gait velocity and
variability in older adults with mild cognitive impairment. Journal of Exercise
Rehabilitation, 2018. 14(6): p. 993-997.
44. Corp, D.T., et al., Reduced motor cortex inhibition and a ‘cognitive-first’
prioritisation strategy for older adults during dual-tasking. Experimental
Gerontology, 2018. 113: p. 95-105.
45. Booth, V., V. Hood, and F. Kearney, Interventions incorporating physical and
cognitive elements to reduce falls risk in cognitively impaired older adults: A
systematic review. JBI Database of Systematic Reviews and Implementation Reports,
2016. 14(5): p. 110-135.
46. Plummer, P., et al., Effects of Physical Exercise Interventions on Gait-Related Dual-
Task Interference in Older Adults: A Systematic Review and Meta-Analysis.
Gerontology, 2015. 62(1): p. 94-117.
47. Wajda, D.A., et al., Intervention modalities for targeting cognitive-motor interference
in individuals with neurodegenerative disease: a systematic review. Expert Review
of Neurotherapeutics, 2017. 17(3): p. 251-261.
48. Modarresi, S., et al., Gait parameters and characteristics associated with increased
risk of falls in people with dementia: a systematic review. International
Psychogeriatrics, 2018: p. 1-17.
-14-
CHAPTER 2: SYSTEMATIC REVIEW
The Effects of Dual-Task Walking on Gait Dynamics in Older Adults with Cognitive
Impairment: A Systematic Review
Authors and contributions:
Tess C Hawkins MExPhys, Maria A Fiatarone Singh MD, Trinidad Valenzuela Arteaga PhD,
Jeffrey M Hausdorff PhD, and Yorgi Mavros PhD.
TH, YM, and MFS contributed to the conception of the work. All authors contributed to the
acquisition, analysis, or interpretation of data. TH drafted the manuscript and all authors
critically revised the manuscript. All authors gave final approval and agree to be accountable for
all aspects of work ensuring integrity and accuracy.
-15-
Faculty of Health Sciences
2.1 AUTHOR CONTRIBUTION STATEMENT
Candidate Name: Tess C Hawkins
Degree Title: Master of Applied Science (Research)
Paper Title: The effects of dual-task walking on gait dynamics in older adults with cognitive
impairment: A systematic review
As the research supervisor of the above candidate, I confirm that the above candidate has made
the following contributions to the above paper title:
- Conception and design of the research
- Analysis and interpretation of the findings
- Writing the paper and critical appraisal of content
Professor Maria Fiatarone Singh
Discipline of Exercise & Sport Science
Faculty of Health Sciences
The University of Sydney
-16-
2.2 ABSTRACT
Objectives
Cognitively impaired individuals have greater variability in gait dynamics than cognitively
healthy individuals. The stress of dual-tasking reveals abnormalities in gait dynamics not
observed under single-task conditions, and may be particularly relevant in individuals with
cognitive impairment. We aimed to review the cost of dual-tasking on gait dynamics in
cognitively impaired older adults, according to cognitive function diagnosis, type of dual-task
paradigm used and/or method of gait dynamics measurement, as well as to identify any clinical
characteristics associated with dual-task costs.
Method
This systematic review was conducted according to the PRISMA guidelines and prospectively
registered in the PROSPERO database (no. CRD42018105787). An electronic database search
was conducted on the 9th August 2018 in the following databases; Ageline, CINAHL, EBM
review database CCRCTs, EMBASE, Medline, PEDro, PsychINFO, Scopus and Web of
Science. An email alert system for new published articles was set up, with the last record from
this alert system screened on the 15th February 2019. A random-effects meta-analysis was
performed when I2 was <75%, alternatively, a narrative synthesis was undertaken.
Results
Among 16,519 citations, 25 articles met the inclusion criteria. All studies included a single-
and dual-task walking condition for a cognitively impaired group [Mild Cognitive Impairment
(MCI) or dementia]. Fourteen studies included a cognitively healthy comparison group and
-17-
3studies included both an MCI group and a dementia group. Twenty-seven different
spatiotemporal and nonlinear measures of gait dynamics and 20 different dual-task procedures
were identified in this literature. Gait dynamics in cognitively impaired older adults are
worsened under dual-task conditions compared to single-task conditions. In individuals with
cognitive impairment, the dual-task cost is higher than it is in healthy older adults. A meta-
analysis included the 3 studies that allowed for a direct comparison between adults with
dementia and MCI, and showed a significantly greater dual-task gait cost in dementia vs. MCI,
with a relative effect size of 0.60 [0.22, 0.99].
Conclusions
Cognitive disease severity increases as gait dynamics become more impaired. Research is
lacking into nonlinear gait dynamics, the relationship to fall risk, and other characteristics which
may be associated dual-task gait dynamics. More well-designed longitudinal studies and
controlled trials with adequately powered samples are needed to confirm the clinical utility and
predictive value of dual-task gait testing, as well as to provide a consensus on the most robust
methods of dual-tasking and gait outcome assessment.
Key words
Dual-task, Gait variability, Cognitive impairment, MCI, Dementia.
-18-
2.3 INTRODUCTION
Consistent, safe walking is essential for older adults to maintain independent living and avoid
falls [1]. Falls are a global public health concern, due to the increasing number of older people
world-wide [2]. One-third of adults over 65 years of age fall each year [3], resulting in an annual,
overall cost of more than $600 million in Australia [4]. Although falls are multifactorial, stride-
to-stride fluctuations within an individual’s walking pattern (also referred to as gait dynamics)
[5] reduce the stability of gait and the ability to resist perturbations or stressors leading to falls.
One such stressor is dual-tasking, which refers to the performance of two tasks simultaneously,
such as walking and talking. Dual-task paradigms are used in research to recreate situations
where adults are at increased risk of falling [6]. In particular, dual-task walking has been
investigated as a potential marker of mobility and falls risk in individuals with cognitive
impairment [7], where the additional task acts as a cognitive stressor, diverting attention away
from stable locomotion in those who have fewer cognitive reserves to cope with such stressors
[8]. Gait dynamics are worsened under these dual-task conditions, with future fall risk predicted
by the magnitude of gait instability induced by dual-tasking [9].
Thus, the gait dynamic response to dual-tasking has been used to explore the relationships
between motor and cognitive function. For example, dual-task gait dynamics are worse in
cognitively impaired older adults than in their healthy counterparts [10, 11]. This dual-task
deficit, called ‘dual-task cost’, has been identified as a motor characteristic of cognitive
impairment, specifically mild cognitive impairment (MCI) [8]. Mild cognitive impairment
reflects a transitional state between normal aging and AD [12], and can be further described as
amnestic or non-amnestic [13]. Given the doubled risk of falling in adults with cognitive
-19-
impairment compared to healthy adults [14, 15], it is important to understand whether
performance of complex tasks, involving simultaneous targeting of different aspects of motor
control and cognition, is a major mediator of this increased fall risk. Despite the reported
associations [10] between dual-task cost and falls risk in adults with cognitive impairment, there
are limitations and inconsistencies within the literature. Previous reviews have summarized the
effects of dual-tasking on gait dynamics in neurological populations [16-18], attempted to
determine if dual-tasking can help discriminate adults with MCI from cognitively healthy adults
[19] and adults with dementia [20], as well as the association of dual-task walking on risk of on
falls [21]. However, no review to our knowledge has solely focused on the effects of dual-tasking
on linear and nonlinear gait dynamics in cognitively impaired populations free from other
neurological diseases.
Therefore, the purpose of this review was to investigate the effects of dual-task walking on gait
dynamics in cognitively impaired adults, with detailed assessment of the different gait dynamics
measures, dual-task conditions and levels of cognitive impairment. Our objectives were: 1) to
evaluate dual-task costs in older adults with cognitive impairment, 2) to investigate differences
in cost attributable to cognitive function diagnosis, type of dual-task paradigm used and/or
method of gait dynamics measurement, and 3) to determine physiological and clinical
characteristics related to dual-task cost in cognitively impaired older adults. To our knowledge
this is the first systematic review to investigate the impact of cognitive and/or motor dual-task
walking conditions on gait dynamics measures in adults with cognitive impairment.
2.4 METHODS
-20-
Protocol and registration
The protocol was prospectively registered with PROSPERO under registration number
CRD42018105787 at: http://www.crd.york.ac.uk/PROSPERO/ on 17/09/2018 date.
Eligibility criteria
Studies were eligible if they met the following criteria: (a) participants subjected to both single-
and dual-task walking conditions within a cross-sectional study design, or at baseline for a
longitudinal observational or experimental study design; (b) full-length article was published in
a peer reviewed journal or an unpublished thesis accessible by reasonable means; (c) human
participants with at least one type of diagnosed objective cognitive impairment, broadly
including mild cognitive impairment (MCI) and dementia, including Alzheimer’s disease (AD),
vascular dementia, frontotemporal dementia, or other variations with the exception of
Parkinson’s Disease or Lewy Body dementia; (d) a cognitive or motor dual-task walking
condition defined as the “simultaneous processing of two (and sometimes more) sources of
information” [22]; (e) a single-task walking condition to allow for isolation of the effects of the
dual-task condition; and (f) one or more objective measures of gait dynamics defined as spatial,
temporal or nonlinear intra-individual stride-to-stride or step-to-step fluctuations in walking
parameters [5].
Studies were excluded if: (a) cognitively impaired participants were grouped with and could not
be separated from participants with: a documented disease with motor effects that had the
potential to impact gait, including but not limited to Parkinson’s disease, Lewy body dementia,
multiple sclerosis (MS), cerebral palsy, spinal cord injury, traumatic brain injury or any other
-21-
disease that has known motor effects on gait; peripheral neuropathy from any cause, including
but not limited to chemotherapy, Charcot-Marie-Tooth (CMT) disease, diabetes mellitus, or
alcoholism; an intellectual disability at birth; cognitive decline due to a delirium or any cause,
psychiatric disorder, medication or other substance use; (b) participants had self-reported
memory concern or a subjective cognitive impairment without a formal diagnosis of cognitive
impairment; (c) participants had undergone one or more above knee amputation (AKA) or below
knee amputation (BKA), or were born without one or part of, their lower limb(s) greater than
that of a single toe; (d) the dual-task gait dynamics outcome of participants with cognitive
impairment was unable to be isolated from participants without cognitive impairment; (e) gait
dynamics outcomes were unable to be extracted from the article. The exclusion of studies of
neurological diseases or musculoskeletal conditions noted above was applied in order to focus
on the cognition or age-related gait dynamics rather than deficits related to these specific
pathologies, which may require different future preventive or therapeutic strategies.
Search strategy
The following electronic databases were selected from the earliest possible date to February
2019: Ageline, CINAHL, EBM review database CCRCTs, EMBASE, Medline, PEDro,
PsychINFO, Scopus and Web of Science. Further, email alerts were set up on all databases and
reviewed weekly until 15th February 2019. To maximize the search sensitivity, the search
strategy included a combination of ‘intervention’ and ‘outcome’ terms, however, it did not
include ‘population’ or ‘comparison’ terms. The intervention terms included ‘dual task*’ OR
‘dual-task*’ OR ‘multi task*’ OR ‘multi-task’ OR ‘secondary task*’ OR ‘attention task*’ OR
‘cognitive task*’ OR ‘motor task*’ OR ‘two task*’ OR ‘2 task*’. The outcome terms included
-22-
walk* OR gait OR locomot* OR ambulat* OR stride* OR step* OR ‘double limb’ OR ‘double-
limb’ OR ‘double support’ OR ‘swing time’ OR ‘stride-to-stride’ OR ‘stride to stride’ OR ‘foot
clearance’ OR mobility OR stability OR instability OR ‘gait variability’ OR ‘centre of pressure’
OR ‘center of pressure’ OR COP OR ‘centre of mass’ OR ‘center of mass’ OR COM OR ataxia
OR McRoberts OR ‘Gait Up’ OR APDM OR GAITRite OR Axivity OR AX3 OR Opal* OR
Pedar OR Zeno OR Gyroscope* OR Lyap* OR fractal. The coding of the search strategy for
each database was customized to search multipurpose (.mp) and the intervention and outcome
searches were then combined with ‘AND’ to produce the final results pool. No language or date
restrictions were applied to the search strategy. A full electronic search strategy example is
presented in Table 2.1. Further eligible trials were hand-searched from the reference lists of all
eligible studies and relevant reviews. Where necessary, authors were contacted for full text
articles or complete gait dynamics datasets in order to identify potential additional studies or
clarify data extracted.
Study selection
One reviewer (TH) performed the literature search and study selection process, which included
the removal of duplicate articles and articles with irrelevant titles or abstracts. The full texts of
the remaining articles were read (TH and TV), with all articles that did not meet the inclusion
criteria removed, and reasoning documented. Any disagreements were resolved by consensus
with other authors (YM and MFS). Remaining eligible articles were included in the systematic
review. In the case that 2 papers were published with the same data set, the paper that was first
published was included [23, 24].
-23-
Data collection process
TH extracted data from each eligible article into pilot-tested collection tables. The data were
verified by direct comparison to the original article (TV). Any discrepancies or disagreements
in data were reviewed and resolved by consensus prior to tabulation (MFS). The data extraction
tables were subsequently refined for increased readability for final manuscript publication. In
the case that multiple walking trials for the same walking condition were conducted, the data
that were collected first were included for analysis to be consistent with studies that included
only 1 trial [25]. In the case of a longitudinal trial with repeated testing, only the baseline data
were used for analysis [26, 27]. In the case that there was more than 1 “healthy” comparison
group, the group that was most closely matched in clinical characteristics (e.g., age, clinical
status) to the cognitive impairment group was included for analysis [28, 29].
Risk of bias assessment
Risk of bias and quality of the included articles were independently appraised by 2 reviewers
(TH and TV) using a modification of the 27-point Downs and Black checklist [30]. Appraisal
was based on reporting, internal validity (bias and confounding) and external validity. Papers in
this review were likely to investigate dual-task walking as an acute exposure. Due to the nature
of this review, items related to follow up (items 9, 17, 26) were not relevant and not included for
scoring. Additionally, item numbers 5, relating to confounder distribution, and 27, relating to
statistical power, were modified to be consistent with the scoring of other items (i.e., alteration
from a 0 to 2 scale, and a 0 to 5 scale, to a 'no, 0; unable to determine, 0; and yes, 1'). Criterion
23 was altered for increased specificity to read: ‘Was the order of the walking tasks (single and
dual) randomized for study subjects?’. To determine if studies were sufficiently powered, a
-24-
clinically significant dual-task cost was defined with reference to the paper by Springer et al.
[31] which reported the difference in dual-task cost between fallers and non-fallers. Therefore,
the maximum modified Downs and Black score possible was 24, with a higher score indicating
better quality.
Summary measures
Primary outcome measures included the objective measurement of spatial, temporal or nonlinear
measures of gait dynamics of stride width, stride time, foot clearance, swing time, stance time,
inconsistency of variance and other aspects of gait dynamics. Gait dynamics could be measured
by any device that recorded intra-individual stride-to-stride fluctuations in walking parameters
[5]. Outcomes may have been expressed as linear measures of gait dynamics such as a co-
efficient of variation (CV) or standard deviation (SD) of stride-to-stride or step-to-step
fluctuations, or nonlinear measures such as approximate entropy (ApEn) or detrended fluctuation
analysis (DFA), which quantify gait irregularity and unpredictability over time-series data [32].
Adverse events related to the dual-task paradigm were extracted if reported, such as falls.
Synthesis of results and Analyses
Studies were split into 3 groups; (a) studies without a control comparison group; (b) studies that
allowed for a comparison between adults with MCI or dementia with adults with a cognitively
healthy control comparison group; and (c) studies that allowed for a direct comparison between
adults with dementia and adults with MCI.
-25-
The primary outcome of interest for this review was gait dynamics as defined above. Gait speed,
number of new falls in the past 12 months (or any shorter time frame) and number of injurious
falls in the past 12 months (or any shorter time frame), visual impairment, hearing impairment,
orthostasis, physical fitness such as aerobic capacity, strength, balance or functional capacity,
depressive symptoms, self-efficacy, nutritional status and quality of life were extracted as
secondary outcomes if reported.
The main data extracted were any quantifiable effects noted on the primary outcome of gait
dynamics, including means and standard deviations (SD) or other summary statistics as
appropriate to the data. The data extracted were at the aggregate level of each study. In addition
to extraction of means and standard deviations (SD) for gait dynamics outcomes, mean
differences (MD) and 95% confidence intervals (95% CI) were calculated and effects sizes (ES)
were calculated as standardized mean difference (SMD) and 95% CI. For studies without a
comparison group, the MD was calculated by subtracting the mean in the single-task condition
from the mean the dual-task condition. The SMD was then calculated by dividing this MD by
the single-task SD using Review Manager 5.3 software (RevMan, Version 5.3; The Nordic
Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark). For studies that had a
comparison group, the MD was calculated by subtracting the mean change from single-to-dual-
task in the control condition from the mean change in single-to-dual-task in the cognitive
impairment condition. The SMD for studies with comparison groups was then calculated by
dividing by the pooled single-task SD. All SMD’s were adjusted for small sample bias (Hedges’
g) [33]. For Tables 2.5, 2.6 and 2.7 the signs of the ESs were reversed in some cases so that a
positive value represented a worsening of gait dynamic measure. Effect sizes were categorized
-26-
according to Cohen’s interpretation of ‘trivial’ (<0.20), ‘small’ (>2.0 to <0.50), ‘moderate’
(>0.50 to <0.80) and ‘large’ (>0.80) [34].
Random effects meta-analyses were attempted for all measures of gait dynamics, with ESs
pooled when I2 was less than 75% using Review Manager (RevMan, Version 5.3; The Nordic
Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark). The I2 range was
chosen to reflect the Cochrane Handbook for Systematic Reviews of Interventions [35]
interpretation of heterogeneity, where an I2 measures of 75-100% reflects ‘considerable
heterogeneity’. For the purposes of the meta-analyses only, when multiple outcomes of gait
dynamics where reported for a group of participants, or a single group had multiple comparisons
to other groups, the total sample size for that group divided by the number of
outcomes/comparisons for that group to account for dependency was used, as recommended by
Cochrane [35]. First, meta-analyses were attempted within the 3 groups (i.e., single- vs. dual-
tasking in MCI or dementia with no cognitively intact control comparison group; cognitively
impaired adults compared to cognitively health adults; and adults with MCI compared to adults
with dementia). Only the direct comparison between adults with MCI and adults with dementia
showing sufficient homogeneity to be pooled (I2= 0%). Attempts to reduce heterogeneity
sufficiently (I2<75%) in the other two groups were ultimately unsuccessful. These steps included
stratifying by cognitive status (MCI or dementia), followed by selecting the most commonly
used measurement of gait dynamics (stride time CV), and finally by stratifying by the cognitive
task. Studies were then removed, and the change in I2 and Q noted, with the study resulting in
the largest reduction removed. As the resultant I2 was not <75%, then this same process was
repeated and a second study removed. However, the I2 still showed considerable heterogeneity
-27-
(I2>75%), and so a meta-analysis was not performed, and a narrative review of the results from
these groups is provided. Results were separated by linear and nonlinear methods of assessing
gait dynamics. Due to the high number of studies reporting coefficient of variation in stride time
as an outcome, stride time CV was reported separately from other linear measures of gait
dynamics, and the MD was reported together with ES for ease of interpretation of clinical
meaningfulness of results. Otherwise, ES was used when reporting studies that used different
outcomes (e.g., swing time CV and stride regularity). Summary results in the narrative synthesis
are presented as MD (range) and ES (range).
2.5 RESULTS
Study selection
The initial keyword search returned 16, 519 results. Following the removal of duplicates, and
title and abstract exclusions, 49 full-text articles were evaluated (Figure 2.1). A further 25 articles
were excluded on the basis of eligibility. In 4 studies where complete data were unable to be
extracted the corresponding authors were contacted. No response was received from 3 studies
[36-38] (n = 3), with a response from 1 study [39], which was then included, for a total of 25
eligible studies.
Study design
In the 25 studies included there were 3 distinct study designs: cross-sectional, longitudinal
observational and longitudinal experimental. This included 10 studies with more than 1 cognitive
impairment comparison group, 9 studies with more than 1 dual-task procedure and 9 studies with
-28-
more than 1 gait dynamics outcome measured. The cohort characteristics of the included studies
are displayed in Table 2.2.
Quality
Modified Downs and Black scores are shown in Table 2.3. The average study quality was
moderate, 13.6 (range: 11 – 17/24). The most common limitations were lack of subject and
assessor blinding, adverse event reporting, reporting of participant representativeness within
population and sample, recruited participant source population and representativeness,
recruitment time frame, and concealment of intervention/procedure. It is acknowledged that
blinding of participants and assessors, and the concealment of intervention/procedure to
participants were not possible due to the study designs, thus potentially limiting the maximal
score to 21 rather than 24. Additionally, as most study designs were acute exposure, the
opportunity for an adverse event to occur was reduced, which potentially limited the reporting
of such events. Lack of reporting of participant representativeness within the population and
sample, however, are threats to external study validity and were deficient in most studies.
Cohort characteristics
Cohort characteristics are presented in Table 2.2. Across all studies 1118 participants (56.5%
female) were included, 797 cognitively impaired (55.2% female) and 321 cognitively healthy
(59.8% female). Twenty-three studies (92.0%) included both male and female participants and
2 studies (8.0%) [40, 41] did not specify sex. The mean reported age for all included study
participants was 76±5 years (range: 59-94 years), with the mean reported age for the cognitively
impaired 77±6 years (range: 61-92 years) and for the cognitively heathy 74±4 years (range: 59-
-29-
94 years). Habitual gait speed was reported in 19 studies [8, 11, 23, 25-29, 39-49] for cognitively
impaired participants (0.94±0.25 m/s) and 9 studies [11, 23, 28, 29, 40, 42, 45-47] for cognitively
healthy participants (1.11±0.16 m/s).
Studies included participants with varying cognitive impairment diagnoses: 14 studies (52.0%)
[25, 26, 28, 41-43, 45, 46, 48-53] included dementia only, 8 studies (32.0%) [8, 10, 27, 29, 39,
40, 44, 47] included MCI only and 3 studies (12.0%) [11, 23, 54] included both dementia and
MCI. Fourteen studies (56.0%) [10, 11, 23, 28, 29, 40, 42, 45-47, 50, 52-54] included a
cognitively healthy control group as a comparison group. The Winblad [55] criteria and the
Petersen [12] criteria were the most common MCI diagnostic criteria used, cited by 4 [10, 11,
27, 29] and 3 studies [8, 44, 54], respectively. The National Institute of Neurological and
Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders
Association (NINCDS-ADRDA) [56] and the Diagnostic and Statistical Manual of Mental
Disorders (DSM-IV) (4th edition) were the most common dementia diagnostic criteria, cited by
6 [23, 26, 41-43, 49] and 3 studies [26, 42, 50], respectively. Numerous assessments were used
to determine the severity of MCI and dementia, with the Mini-mental State Exam (MMSE) the
most common. Twenty-two studies [8, 10, 11, 23, 25, 26, 28, 39-46, 48-54] reported MMSE
scores for cognitively impaired participants (21.6±4.1), specifically MCI (26.4±1.1) and
dementia participants (19.9±0.9), and 12 studies [10, 11, 23, 28, 40, 42, 45, 46, 50, 52-54]
reported MMSE scores for cognitively healthy participants (28.8±1.1).
Measurement of gait dynamics
Characteristics of dual-task procedures are presented in Table 2.4. All studies included flat
ground walking and participants were instructed to walk at their usual pace. The mean measured
-30-
walking distance was 21.66m (range: 3.66-160m). Twelve studies [8, 10, 11, 25-27, 43, 48-52]
used the GAITRite system (CIR Systems Inc., Clifton, NJ, USA) to measure gait dynamics, 4
studies [41, 42, 47, 52] used footswitch sensors, 3 studies [23, 39, 44] used Locometrix (Centaure
Metrix, Evry, Essonne, France), 2 studies [28, 46] used DynaPort MiniMod (McRoberts BV,
The Hague, The Netherlands), 2 studies [40, 53] used motion capture cameras and 3 studies [29,
45, 54] used movement tracking devices. The mean number of trials per walking condition was
1.7 (range: 1-6 trials), with 2 studies [23, 43] collecting data from 1 pre-nominated trial when
multiple trials were completed per condition, (e.g., data were collected for the second trial out of
the 3 conducted trials per condition), and 7 studies [8, 10, 11, 29, 41, 45, 50] not reporting the
number of trials completed. The order of single-task and dual-task walking was inconsistent
between studies; 9 studies [8, 10, 11, 27, 42, 47, 50, 51, 53] completed the tasks in a random
order, 7 studies [25, 39-41, 43, 44, 49] completed the tasks in a set order, (i.e., single-task then
dual-task), and 9 studies [23, 26, 28, 29, 45, 46, 48, 52, 54] did not report the task order. No
studies with a non-randomized task order corrected their analytical models for order effects.
Dual-task procedure characteristics
Characteristics of dual-task procedures are presented in Table 2.4. Dual-tasks were categorized
into two types: cognitive and motor. Twenty-four studies implemented cognitive protocols [8,
10, 11, 23, 25-29, 39-48, 50-54] and 2 studies implemented motor protocols [48, 49]. The
cognitive protocols were further separated into 4 sub-types [16]: mental tracking, verbal fluency,
working memory and verbal memory. Mental tracking, the task of holding information mentally
while manipulating the same information, to measure sustained attention [57], was carried out
in 21 studies [8, 10, 11, 23, 26, 27, 29, 39, 40, 42, 44, 45, 47, 48, 50-54]. All 21 studies reported
-31-
backwards counting, however, the protocol varied between studies with the starting number
ranging from 30 to 378 and the counting increment ranging from 1 to 7. The most common
protocols were backward counting from 50 and 100 by 1s, carried out by 7 [23, 26, 39, 42, 44,
50, 52] and 5 studies [8, 10, 11, 29, 45], respectively. Verbal fluency, the task of spontaneously
producing words within specific constraints to measure executive function [57], was carried out
in 12 studies [8, 10, 11, 25, 27-29, 43, 46, 47, 51, 53]. Two verbal fluency protocols were
reported: animal naming and categorical letter naming, which 5 studies [8, 10, 11, 27, 29] and 3
studies [28, 46, 47] carried out. Working memory, the task of holding information mentally for
later processing [58] was carried out in 5 studies [25, 41, 43, 51, 53] and included forwards
counting [25, 43, 51] and forward digit span [41, 53] protocols. Forwards counting and forward
digit span were included as working memory tasks, rather than mental tracking tasks, as forward
recall loads onto a separable short-term memory factor unlike backward counting and digit span,
which require a more attention-demanding transformation of the digit sequence [59]. Verbal
memory, the task of recalling specific past events or information using speech [60], was carried
out in 1 study [47] in the form of short story recall.
Gait dynamics outcomes
Gait dynamics outcomes are presented in Table 2.5, 2.6 and 2.7. Twenty-six spatiotemporal gait
dynamics outcome measures were examined across all studies. Stride time CV, step length CV
and stride regularity were the most commonly measured, with 18 studies [8, 10, 11, 25-28, 40-
42, 45-47, 49-53], 4 studies [25, 48, 49, 53] and 3 studies [29, 39, 44] reporting each outcome,
respectively.
-32-
Single- vs. dual-tasking in MCI or dementia with no cognitively intact control comparison group
Among the 25 studies, 11 studies had at least 1 cohort with MCI and 17 studies had at least 1
cohort with dementia, resulting in 84 ESs comparing single- to dual-tasking (Table 2.8). A meta-
analysis was attempted but was not appropriate due to considerable heterogeneity (I2=85%,
Tau2=1.32, Q=553.30, p<0.0001). Attempts to reduce heterogeneity included stratifying by
cognitive status (MCI or dementia), and then further stratifying by gait dynamics outcome and
cognitive task, but were ultimately unsuccessful. Therefore, a narrative synthesis of ESs is
presented.
Stride time CV
In cohorts with MCI, stride time CV was the most commonly used outcome of gait dynamics,
resulting in 17 ESs from 6 studies (Table 2.5a). Ten out of the 17 ESs involved the addition of a
mental tracking cognitive dual-task (5 counting backwards by 1; 1 counting backward by 2; and
4 counting backwards by 7), 6 used a categorical verbal fluency dual-task (5 used animal naming
and 1 used letter specific word naming), while the remaining ES was verbal memory in the form
of a short story recall. Overall, the data showed that the addition of a cognitive dual-task resulted
in a significant increase in stride time CV, with 15 of 17 ESs reported being significant (MD
ranging from 1.42% to 8.77%). The 2 non-significant results were from the same study [8], which
used backwards counting by 1, with the subgroup with non-amnestic MCI unexpectedly
achieving a smaller dual-task cost with the addition of a cognitive distractor. Overall, the median
MD was 2.43% (range: 0.50% to 8.77%), while ES was 1.87 (range: 0.35 to 7.70).
-33-
Similarly, in cohorts with dementia, stride time CV was the most commonly used measure of
gait dynamics, resulting in 20 ESs from 13 studies (Table 2.5b). Eleven out of the 20 ESs
involved the addition of a mental tracking cognitive dual-task (8 counting backwards by 1; 1
counting backward by 7; 1 counting backwards by an unspecified number; and 1 backward 3-
digit span), 5 used a working memory cognitive task (2 counting forwards by 1; 1 counting
forwards by an unspecified number and 2 forward 3-digit span), 3 used a categorical verbal
fluency dual-task (1 used animal naming and 2 used letter specific word naming), while the
remaining ES was a motor dual-task of tray carrying. Overall, the data showed that the addition
of a cognitive dual-task resulted in heterogeneous effects on stride time CV, with 12 significant
(MD ranging from 2.19 to 29.00) and 7 non-significant ESs (MD ranging from -0.53% to 3.00%).
Three non-significant studies [28, 41, 52] still reported potentially clinically meaningful MD
changes in dual-task cost (MDs of 2.6%, 2.6% and 3.0% for stride time CV), suggesting the
possibility of type II error, with respective ESs of 0.74 (95% CI -0.03, 1.51), 0.97 (95% CI -0.16,
2.10) and 0.53 (95% CI -0.50, 1.57). Overall, the median MD was 2.97% (range: -0.53% to
29.00), while ES was 1.26 (range: -0.18 to 42.18).
Other linear measures of gait dynamics
In cohorts with MCI, 11 ESs from 5 studies were calculated using other linear measures of gait
dynamics (Table 2.5a). These included step regularity (n=3), stride regularity (n=5), step time
CV (n=2) and step time variance (n=1). Results were heterogeneous with 6 significant ES
(ranging from 1.03 to 26.65). Overall, the median ES for all other linear measures of gait
dynamics was 1.07 (range: 0.24 to 2.74).
-34-
In cohorts with dementia, 14 other methods of linear gait dynamics were used across 9 studies,
resulting in 34 ESs (Table 2.5b). Twenty-nine of these ESs were generated from relatively simple
cognitive dual-tasks that involved either forward counting by 1 (n=16), backward counting by 1
(n=11), forward 3-digit span (n=1) or backward 3-digit span (n=1). Results were mostly
negative, suggesting little difference between single- and dual-task gait dynamics, with 22 of 34
ESs non-significant, and the median ES for all other linear measures of gait dynamics 0.51
(range: -0.40 to 45.71).
Nonlinear measurements of gait dynamics
Only 2 studies used nonlinear measures of gait dynamics, resulting in 3 ESs. Gait dynamic
outcomes for cohorts with MCI and dementia are presented in Tables 2.5a and 2.5b, respectively.
One study was in adults with MCI (2 ESs), while the remaining study was in adults with
dementia. In adults with MCI, gait dynamics were assessed using ApEn using backwards
counting by 1 and animal naming as the cognitive distractors, with both resulting in very large,
significant worsening of gait [ES = 14.93 (95% CI 7.05, 22.81) for backwards count by 1, and
ES = 26.97 (95% CI 12.85, 41.09) for animal naming]. By contrast, in adults with dementia, gait
dynamics were measured using DFA, with a word naming verbal fluency task, showing no
change in gait dynamics during the dual-task [ES = 0 (95% CI -0.19, 0.19)].
Cognitive impairment vs. control comparison group
Fourteen studies included a healthy control group; 7 MCI studies [10, 11, 23, 29, 40, 47, 54] and
10 dementia studies [11, 23, 28, 42, 45, 46, 50, 52-54] resulting in 40 ESs. A meta-analysis was
attempted but was not appropriate due to considerable heterogeneity (I2=100%, Tau2=3.35,
-35-
Q=20054.93, p<0.00001). Attempts to reduce heterogeneity were unsuccessful, which included
stratifying by cognitive status (MCI or Dementia), as well as stratifying by gait dynamics
outcome and cognitive task, results summarized in Table 2.8). Therefore, a narrative synthesis
of ESs is presented.
Stride time CV
In cohorts with MCI, stride time CV was the most commonly used outcome of gait dynamics,
resulting in 9 effect sizes from 4 studies. Five out of the 9 ESs involved the addition of a mental
tracking cognitive dual-task (3 counting backwards by 1; 2 counting backwards by 7), 3 used a
categorical verbal fluency dual-task (2 used animal naming; 1 used letter specific word naming),
while the remaining ES was verbal memory in the form of a short story recall. Overall, the data
showed that the dual-task cost of a cognitive dual-task for adults with MCI compared to
cognitively healthy controls resulted in a significant increase in stride time CV, with 8 of 9 ESs
reported significant (MD ranging from 1.07% to 7.05%). Gait dynamic outcomes for cohorts
with MCI compared to cognitively healthy controls are presented in Table 2.6a. Overall, the
median MD was 2.75% (range: 0.26% to 7.05%) and ES 2.43 (range: 0.16 to 6.53).
Similarly, in cohorts with dementia compared to cognitively healthy controls, stride time CV
was the most commonly used measure of gait dynamics, resulting in 13 ESs from 8 studies. Nine
out of the 13 ESs involved the addition of a mental tracking cognitive dual-task (7 counting
backwards by 1, 1 counting backward by 7, and 1 backward 3-digit span), 3 used a categorical
verbal fluency dual-task (1 used animal naming and 2 used letter-specific word naming), while
the remaining ES was a working memory cognitive task (forward 3-digit span). Overall, the data
-36-
showed that the addition of a cognitive dual-task for adults with dementia compared to
cognitively healthy controls resulted in heterogeneous effects on stride time CV, with 9 ESs
showing worsening of gait (MD ranging from 1.79% to 16.64%), 1 ES showing an improvement
in gait (MD -2.58%) and 3 non-significant ESs (MD ranging from (-0.40% to 2.25%). Gait
dynamic outcomes for cohorts with dementia compared to cognitively healthy controls are
presented in Table 2.6b. Overall, the median MD was 2.23% (range: -2.58% to 16.64%) and ES
1.97 (range: -1.09, 22.03).
Other linear measures of gait dynamics
In cohorts with MCI, 8 effect sizes from 3 studies were calculated for other linear measures of
gait dynamics, including step regularity (n=3), stride regularity (n=2), step time CV (n=2) and
step time variance (n=1). Six of the 8 ESs were significant. Non-significant results used the
simple task of backwards counting by 1 (n=2) only. Gait dynamic outcomes for cohorts with
MCI compared to cognitively healthy controls are presented in Table 2.6a. Overall, the median
ES for all other linear measures of gait dynamics was 0.63 (range: 0.23, 1.36).
In cohorts with dementia, 8 ESs from 4 studies were calculated from linear measures of gait
dynamics. The 8 ESs were generated from cognitive dual-tasks of backwards counting by 1
(n=6), backwards 3-digit span (n=1) or forward 3-digit span (n=1). Results were heterogeneous
with 6 of 8 ESs significant but varying greatly in magnitude (ranging from 0.66 to 23.20). Gait
dynamic outcomes for cohorts with dementia compared to cognitively healthy controls are
presented in Table 2.6b. Overall, the median ES for all other linear measures of gait dynamics
was 1.20 (range: -0.10 to 23.2).
-37-
Nonlinear measurements of gait dynamics
Only 2 studies used nonlinear measures of gait dynamics, resulting in 3 ESs. Gait dynamic
outcomes for cohorts with MCI compared to cognitively healthy controls and dementia
compared to healthy controls are presented in Table 2.6a and Table 2.6b, respectively. Two ESs
came from 1 study [29] in adults with MCI, while the remaining ES was from a study of adults
with dementia [46]. In adults with MCI, gait dynamics were assessed via ApEn using animal
naming and backwards counting by 1 as the cognitive distractors. Neither distractor resulted in
a significant difference in dual-task cost in the MCI cohort [ES = 0.45 (95% CI -0.29 to 1.20)]
and [ES = -0.62 (95% CI -1.37 to 0.14)], respectively. By contrast, in adults with dementia
compared to cognitively healthy controls, gait dynamics were measured using DFA, with a word
naming verbal fluency task, showing a significant worsening of DFA during dual-tasking [ES =
0.81 (95%CI: 0.01, 1.62)].
MCI vs. dementia groups
Three studies included both MCI and dementia groups [11, 23, 54] and measured gait outcomes
using stride time CV [11], step regularity [23] and step time variance [54]. Gait dynamic
outcomes for cohorts with MCI vs. dementia are presented in Table 2.7. A meta-analysis was
performed (Figure 2.2), showing that adults with dementia have a moderate, significant
worsening of gait dynamics under dual-task conditions compared to adults with MCI [ES = 0.60
(95%CI: 0.22, 0.99), I2=0%, Tau2=0.00, Q=3.74, p=0.002).
Adverse events
-38-
Only 1 study [27] reported adverse events, however this study was a placebo-controlled drug
study where none of the reported events were attributed to baseline single- or dual-task
procedures. No study reported whether single- or dual-task walking was associated with any
adverse event such as falling.
Other cohort characteristics
Limited information was available on cohort characteristics other than gait or cognitive
performance, with fall history and measures of physical fitness being the most often reported
characteristics. Six studies [8, 10, 25, 42, 48, 49] reported falls history in the past 6 to 12 months.
A sensitivity analysis adjusting for history of falls and age was conducted in 1 study [10], the
direction and magnitude of the dual-task costs were maintained (a difference <10% from the
unadjusted values). A comparison between multiple fallers and non-multiple fallers was
conducted in 1 study [48], which showed that there was no significant interaction between dual-
task cost and faller status. One study [42] adjusted for previous falls and other baseline
characteristics using a multivariate linear regression for single-task and dual-task conditions and
stride time CV, which showed that previous falls were not significantly related to stride time CV
under these conditions. However, no statistical analysis with respect to dual-task cost of gait
dynamics reported. A statistical analysis by faller status with respect to dual-task cost of gait
dynamics was not reported in 3 other studies [8, 43, 49]. History of falls was listed as an
exclusion criterion in 4 studies [11, 23, 40, 44]. Additionally, 2 studies [8, 10] reported fear of
falling, with no statistically significant between-group comparisons in either study (p=0.77 [10]
and p=0.84 [8]) and no statistical analysis with respect to the dual-task cost of gait dynamics was
reported.
-39-
Physical activity or functional mobility assessments were reported in 6 studies [8, 10, 11, 23, 29,
39], which included self-reported physical activity level, one-legged balance, and Timed Up and
Go (TUG) test. Three studies [8, 10, 11] reported self-reported physical activity level. One study
[8] adjusted for physical activity level using a multivariable linear regression, and, as previously
reported, amnestic MCI participants had statistically significantly higher stride time CV (p=0.01)
compared to non-amnestic participants under single- and dual-task walking after adjustment. No
other studies reported relating these characteristics to gait performance.
Nutritional status was measured in 2 studies [23, 44] using the Mini-Nutritional Assessment
(MNA), however, no statistical analysis with respect to the dual-task cost of gait dynamics was
reported. Five studies [23, 26, 27, 39, 44] reported that depression was an exclusion criterion,
measured by the Geriatric Depression Scale (GDS) [61] or the Hospital Anxiety Depression
Scale (HADS) [62], but these scores were not included in models of gait dynamics. Injurious
falls in the past 12 months (or any shorter time frame), orthostasis, self-efficacy and quality of
life were not reported in any study.
2.6 DISCUSSION
To our knowledge, this is the first systematic review to solely investigate the effects of dual-task
walking on gait dynamics in older adults with cognitive impairment. The purpose of the review
was to evaluate the effects of dual-task walking on changes in gait dynamics and dynamics with
respect to pathology diagnosis, dual-task paradigm and outcome measure in older adults with
cognitive impairment. In total, 27 different spatial, temporal and nonlinear measures of gait
-40-
dynamics and 20 different dual-task procedures were identified in the literature. The overall
findings of this review are: 1) gait dynamics are worse under dual-task conditions than single-
task conditions in older adults with MCI and dementia; 2) this dual-task cost is greater in
cognitively impaired older adults than in healthy older adults; 3) when MCI and dementia are
directly compared, dual-task cost is greater in older adults with dementia; and 4) characteristics
associated with the dual-task cost of gait dynamics in cognitively impaired older adults have
been minimally investigated.
This review indicates that the addition of a dual-task while walking alters gait dynamics in older
adults with cognitive impairment. This is in agreement with previous studies in healthy older
adults [63], Parkinson’s disease [64], multiple sclerosis [65] and MCI [19]. As expected, dual-
task gait dynamics, measured by stride time CV, increased during mental tracking tasks and
verbal fluency tasks in both older adults with MCI and dementia. This directly reflects outcomes
in healthy older adults, where significant increases in dual-task compared to single-task stride
time CV have been reported for both backwards counting and verbal fluency tasks [66]. As
observed, a more complex task (i.e., dual-tasking compared to single-tasking), produced more
cognitive interference, which resulted in larger variability and a greater cost to gait dynamics
[10, 49]. This increase in dual-task cost is concordant with the literature, which shows that dual-
tasking predicts falls risk better than single-task gait measurements in healthy older adults,
however it has not been as clear in cognitively impaired older adults [9, 67]. If ways are identified
to minimize the cost of dual-tasking on gait dynamics, the frequency and severity of future falls
could potentially be reduced. However, very little data are available to indicate what the
potentially modifiable contributants to dual-task costs are in cognitively impaired cohorts. We
-41-
searched for potential candidates such as low physical fitness, depression, low physical activity
levels, undernutrition, or fear of falling, but unfortunately there was minimal reporting of any
such characteristics, nor of their effect on dual-task cost. Future studies should include such
characterization, and assess relationships of relevant factors to gait outcomes under single- and
dual-task conditions to advance this field.
The dual-task cost of gait dynamics was larger in cognitively impaired older adults than in
cognitively healthy older adults. Specifically, a greater decrement in dual-task gait dynamics was
observed during mental tracking and verbal fluency tasks in both older adults with MCI and
dementia compared to healthy controls. Gait dynamics are maintained and stable with age in
healthy older adults [68], despite dual-tasking, which differs from the significant worsening
under dual-task conditions reported in cognitively impaired older adults [11], and confirmed by
this review. Physiological and pathological aging impact gait ability and cognitive function, and
the association between these two factors suggests that a complex age-related relationship exists
[66]. Broadly, poor gait performance has been identified as predictive of dementia, with a
stronger association in non-AD dementias than in AD [69]. Specifically, increases in stride time
CV under dual-task conditions have been shown to predict cognitive decline [66] suggesting that
loss of gait stability may be an early sign of brain pathology. Cognitive impairment and executive
function impairment are both associated with an increased risk of falls, while global measures of
cognition are associated with serious fall-related injury [14]. Thus, dual-task gait testing has been
recommended as part of the assessment battery to determine risk of falls in all cognitively
impaired older adults [11]. In addition, it may identify individuals with subtle cognitive changes
-42-
who may benefit from detailed evaluation of cognition and assessment for potentially treatable
aetiologies.
The dual-task decrement was larger for individuals with greater cognitive impairment, i.e.,
dementia compared to MCI. A meta-analysis was performed on the 3 studies [11, 23, 54] that
included both MCI and dementia groups. Individuals with dementia had a significantly and
moderately increased dual-task cost compared to individuals with MCI. Previous studies,
including those analyzed, have reported conflicting results as to whether there is a significant
difference between dual-task gait performance in older adults with MCI and AD [11, 23, 24, 54,
70]. Evidence suggests that older adults with AD are slower on basic mobility tests [70], have
slower gait speed [71], perform worse on cognitive tasks [72], while it has been previously shown
that adults with AD and dementia demonstrate increased gait dynamics during single-task
walking than older adults with MCI [71], and thus a ceiling effect may explain some of the
heterogeneity between studies. Additionally, older adults who are more impaired may potentially
be limited in their ability to adopt protective strategies during dual-tasking (e.g., increasing step
length to compensate for gait abnormalities or dysfunction) [73], however, few comparative
studies exist that investigate potential changes across the cognitive spectrum [11], and the gait
measures within the 3 included studies [11, 23, 54] were inconsistent. When clinically assessing
dual-task gait deficits it is important to consider individual characteristics such as the severity of
motor and cognitive impairments, concurrent tasks complexity and the environmental challenge
on the falls risk [73]. To understand more about the impact of the dual-task on gait dynamics
across the cognitive continuum, additional studies are required to examine gait using comparable
outcome measures and dual-tasks.
-43-
The included studies were too inconsistent in their methodology and reporting to determine if a
dual-task type or a specific gait dynamic measure was better at producing or detecting dual-task
cost. Additionally, only two studies measured gait dynamics using nonlinear outcomes (i.e., DFA
[46] and ApEn [29]). The outcomes for DFA and ApEn are not directly comparable, with the
two outcomes being measured in different cohorts using different dual-task conditions.
Additionally, it is not expected to observe the same outcome for different nonlinear outcomes,
with DFA measuring the degree of randomness in highly non-stationary data and ApEn
measuring the likelihood that a template pattern repeats in a time series [74]. In order to have a
complete understanding of the dynamics of gait, both linear and nonlinear measures are needed.
Further research is required using linear and nonlinear outcomes as the primary measure of
interest, with respect to dual-task walking conditions of varying complexity, in order to enhance
the understanding of falls risk and translation of findings into research and clinical guidelines.
No specific characteristics related to the dual-task gait dynamics in cognitively impaired older
adults were identified. However, the reporting of any cohort characteristics relating to falls were
poor, with few studies documenting history of falls, fear of falling or tracking falls over time. In
these studies, individuals who had fallen more times were likely to have worsened dual-task gait
dynamics than those who had fallen fewer times, however, this did not translate into a
significantly increased dual-task cost [43, 48]. Dual-task gait changes are associated with future
fall risk [9], however, the link between future risk of falls and increased gait dynamics is poorly
studied in cognitive impairment. Previous research has shown a greater dual-task cost of gait
dynamics is associated with progressive cognitive decline and an increase in falls risk, although
-44-
this association was not reported in this review [75]. Few studies reported other characteristics
such as physical activity level static balance, functional mobility, nutritional status and
depression, while no studies reported aerobic capacity, strength, self-efficacy, or quality of life.
If these or other modifiable characteristics were able to be identified, a targeted intervention
could be implemented to potentially reduce the risk of falls in this population [49].
2.7 STRENGTHS
The strengths of this review were that it included a broad, sensitive search strategy across all
years and major databases, resulting in a large number of retrieved citations. Additionally, the
analysis of results was stratified by type of cognitive impairment and dual-task condition to
attempt to create more uniformity in the interpretation of dual-task gait dynamics methodology.
This review does not duplicate previous work and reflects the current literature in this topic area
allowing it to help drive recommendations for knowledge translation and act as a reference point
for future research strategies.
2.8 LIMITATIONS
This review was limited by the use of only one author responsible for the initial study selection
and data extraction. However, full text inclusion and exclusion were independently performed
by two authors, and consensus was obtained by a third author. Additionally, unpublished data
were neither searched for nor included. Furthermore, this review was restricted to one meta-
analysis of three studies due to all other groupings displaying an I2 value greater than 75%. The
I2 value represents the consistency of study results and assesses whether differences in results
between studies are compatible with chance alone. Cochrane categorize an I2>75% as possessing
-45-
considerable heterogeneity, which was used to support the decision of using 75% as a cut off
value. While the small number of studies in the meta-analysis may require caution with regards
to interpretation, the direction of the effect of this analysis was in agreement with all other data
presented within this review. Several attempts were used to reduce the heterogeneity, but were
ultimately unsuccessful. Future studies should attempt to identify factors contributing to
heterogeneity across studies.
2.9 CONCLUSIONS
Gait dynamics worsen under dual-task conditions compared to single-task conditions in
cognitively impaired older adults. The dual-task cost of gait dynamics increases in cognitively
impaired older adults compared to healthy older adults, with worse gait dynamics observed with
a greater degree of cognitive dysfunction. Data are too inconsistent currently to determine which
type of dual-task is best able to expose gait instability, or which measure of gait dynamics best
predicts risk of falls or level of cognitive impairment in older adults. Importantly, only two
studies reported nonlinear gait outcome measures, highlighting an area where more research is
needed to understand the complete impact of dual-tasking on gait dynamics. Additionally, factors
that may impact gait dynamics, including history of falling, fear of falling, physical fitness,
objective physical activity or sedentary behavior, depression, nutritional status, vision, hearing,
overall disease burden and medication use, and quality of life were documented poorly or not at
all. To adequately determine the modifiable and non-modifiable characteristics of dual-task gait
dynamics, more well-designed longitudinal studies and controlled trials with adequately
powered samples are needed. Thus, by identifying characteristics of dual-task gait dynamics,
-46-
clinical interventions could be developed to target these modifiable factors with the aim to reduce
the high falls risk in this population.
-47-
2.10 REFERENCES
1. Callisaya, M.L., et al., Gait, gait variability and the risk of multiple incident falls in older
people: a population-based study. Age and Ageing, 2011. 40(4): p. 481-487.
2. Stewart Williams, J., et al., Prevalence, risk factors and disability associated with fall-
related injury in older adults in low- and middle-incomecountries: results from the WHO
Study on global AGEing and adult health (SAGE). BMC Medicine, 2015. 13(1): p. 147.
3. Lundin-Olsson, L., L. Nyberg, and Y. Gustafson, “Stops walking when talking” as a
predictor of falls in elderly people. The lancet, 1997. 349(9052): p. 617.
4. Bradley, C. and J. Harrison, Hospitalisations due to falls in older people, Australia, 2003-
2004, in Injury research and statistics series no. 32. 2007, AIHW: Canberra.
5. Hausdorff, J.M., Gait variability: methods, modeling and meaning. Journal of
NeuroEngineering and Rehabilitation, 2005. 2(1): p. 19.
6. Hausdorff, J.M., et al., Walking is more like catching than tapping: gait in the elderly as
a complex cognitive task. Experimental Brain Research, 2005. 164(4): p. 541-8.
7. Ayers, E.I., et al., Walking while Talking and Falls in Aging. Gerontology, 2014. 60(2):
p. 108-113.
8. Montero-Odasso, M., et al., The motor signature of mild cognitive impairment: results
from the gait and brain study. Journals of Gerontology Series A-Biological Sciences &
Medical Sciences, 2014. 69(11): p. 1415-1421.
9. Muir-Hunter, S.W. and J.E. Wittwer, Dual-task testing to predict falls in community-
dwelling older adults: a systematic review. Physiotherapy, 2016. 102(1): p. 29-40.
10. Montero-Odasso, M., S.W. Muir, and M. Speechley, Dual-task complexity affects gait in
people with mild cognitive impairment: the interplay between gait variability, dual
tasking, and risk of falls. Archives of Physical Medicine & Rehabilitation, 2012. 93(2):
p. 293-299.
-48-
11. Muir, S.W., et al., Gait assessment in mild cognitive impairment and Alzheimer's disease:
The effect of dual-task challenges across the cognitive spectrum. Gait and Posture., 2012.
12. Petersen, R.C., Mild Cognitive Impairment: Clinical Characterization and Outcome. The
Journal of the American Medical Association, 1999. 281(19): p. 303-308.
13. Petersen, R.C., et al., Current concepts in mild cognitive impairment. Archives of
Neurology, 2001. 58(12): p. 1985-1992.
14. Muir, S.W., K. Gopaul, and M.M. Montero Odasso, The role of cognitive impairment in
fall risk among older adults: a systematic review and meta-analysis. Age and Ageing,
2012. 41(3): p. 299-308.
15. Shaw, F.E., Falls in cognitive impairment and dementia. Clinics in Geriatric Medicine,
2002. 18(2): p. 159-173.
16. Al-Yahya, E., et al., Cognitive motor interference while walking: A systematic review
and meta-analysis. Neuroscience and Biobehavioral Reviews, 2011. 35(3): p. 715-728.
17. Li, K.Z.H., et al., Cognitive Involvement in Balance, Gait and Dual-Tasking in Aging: A
Focused Review From a Neuroscience of Aging Perspective. Frontiers in Neurology,
2018. 9: p. 913.
18. Moon, Y., et al., Gait variability in people with neurological disorders: A systematic
review and meta-analysis. Human Movement Science, 2016. 47: p. 197-208.
19. Bahureksa, L., et al., The Impact of Mild Cognitive Impairment on Gait and Balance: A
Systematic Review and Meta-Analysis of Studies Using Instrumented Assessment.
Gerontology, 2017. 63(1): p. 67-83.
20. Bragatto, V.S.R., et al., Dual-task during gait between elderly with mild cognitive
impairment and Alzheimer: systematic review. Fisioterapia em Movimento, 2017. 30: p.
849-857.
21. Menant, J.C., et al., Single and dual task tests of gait speed are equivalent in the
prediction of falls in older people: a systematic review and meta-analysis. Ageing
Research Reviews, 2014. 16: p. 83-104.
-49-
22. Verhaeghen, P., et al., Aging and Dual-Task Performance: A Meta-Analysis. Psychology
and Aging, 2003. 18(3): p. 443-460.
23. Gillain, S., et al., The value of instrumental gait analysis in elderly healthy, MCI or
Alzheimer's disease subjects and a comparison with other clinical tests used in single
and dual-task conditions. Annals of Physical & Rehabilitation Medicine, 2009. 52(6): p.
453-74.
24. Maquet, D., et al., Gait analysis in elderly adult patients with mild cognitive impairment
and patients with mild Alzheimer's disease: simple versus dual task: a preliminary report.
Clinical Physiology & Functional Imaging, 2010. 30(1): p. 51-56.
25. Camicioli, R., T. Bouchard, and L. Licis, Dual-tasks and walking fast: relationship to
extra-pyramidal signs in advanced Alzheimer disease. Journal of the Neurological
Sciences, 2006. 248(1-2): p. 205-9.
26. Beauchet, O., et al., Gait Changes with Anti-Dementia Drugs: A Prospective, Open-
Label Study Combining Single and Dual Task Assessments in Patients with Alzheimer's
Disease. Drugs & Aging, 2014. 31(5): p. 363-372.
27. Gschwind, Y., et al., Ginkgo biloba special extract LI 1370 improves dual-task walking
in patients with MCI: a randomised, double-blind, placebo-controlled exploratory study.
Aging clinical and experimental research. Issue, 2017. 4: p. 609-619.
28. Ijmker, T. and C.J. Lamoth, Gait and cognition: the relationship between gait stability
and variability with executive function in persons with and without dementia. Gait &
Posture, 2012. 35(1): p. 126-30.
29. Martinez-Ramirez, A., et al., Dual Task Gait Performance in Frail Individuals with and
without Mild Cognitive Impairment. Dementia & Geriatric Cognitive Disorders, 2016.
42(1-2): p. 7-16.
30. Downs, S.H. and N. Black, The feasibility of creating a checklist for the assessment of
the methodological quality both of randomised and non-randomised studies of health
-50-
care interventions. Journal of Epidemiology and Community Health, 1998. 52(6): p. 377-
384.
31. Springer, S., et al., Dual-tasking effects on gait variability: The role of aging, falls, and
executive function. Movement Disorders, 2006. 21(7): p. 950-957.
32. Arif, M., et al., Estimation of the effect of cadence on gait stability in young and elderly
people using approximate entropy technique. Measurement Science Review, 2004. 4(2):
p. 29-40.
33. Hedges, L.V., Distribution theory for Glass's estimator of effect size and related
estimators. journal of Educational Statistics, 1981. 6(2): p. 107-128.
34. Cohen, J., Statistical power analysis for the behavioural sciences. Second edition ed.
1988, Hillside, NJ, USA: Lawrence Erlbaum Associates.
35. Higgins, J. and S.e. Green, Cochrane Handbook for Systematic Reviews of Interventions,
in Version 5.1.0 [updated March 2011]. 2011, The Cochrane Collaboration: The
Cochrane Collaboration.
36. Kikkert, L., et al., Gait characteristics and their discriminative power in geriatric
patients with and without cognitive impairment. Journal of NeuroEngineering and
Rehabilitation, 2017. 14.
37. Montero-Odasso, M., et al., Quantitative gait analysis under dual-task in older people
with mild cognitive impairment: a reliability study. Journal of Neuroengineering &
Rehabilitation, 2009. 6: p. 35.
38. Rucco, R., et al., Spatio-temporal and kinematic gait analysis in patients with
Frontotemporal dementia and Alzheimer's disease through 3D motion capture. Gait &
Posture, 2017. 52: p. 312-317.
39. Auvinet, B., et al., Gait disorders in the elderly and dual task gait analysis: a new
approach for identifying motor phenotypes. Journal of Neuroengineering &
Rehabilitation, 2017. 14(1): p. 7.
-51-
40. Lee, J. and S. Park, Effects of a priority-based dual task on gait velocity and variability
in older adults with mild cognitive impairment. Journal of Exercise Rehabilitation, 2018.
14(6): p. 993-997.
41. Sheridan, P.L., et al., Influence of executive function on locomotor function: divided
attention increases gait variability in Alzheimer's disease. Journal of the American
Geriatrics Society, 2003. 51(11): p. 1633-7.
42. Allali, G., et al., Frontotemporal dementia: pathology of gait? Movement Disorders,
2010. 25(6): p. 731-7.
43. Camicioli, R. and L. Licis, Motor impairment predicts falls in specialized Alzheimer care
units. Alzheimer Disease & Associated Disorders, 2004. 18(4): p. 214-8.
44. Gillain, S., et al., Gait speed or gait variability, which one to use as a marker of risk to
develop Alzheimer disease? A pilot study. Aging-Clinical & Experimental Research,
2016. 28(2): p. 249-255.
45. Hsu, Y.L., et al., Gait and balance analysis for patients with Alzheimer's disease using
an inertial-sensor-based wearable instrument. IEEE Journal of Biomedical & Health
Informatics, 2014. 18(6): p. 1822-30.
46. Lamoth, C.J., et al., Gait stability and variability measures show effects of impaired
cognition and dual tasking in frail people. Journal of Neuroengineering & Rehabilitation,
2011. 8: p. 2.
47. Nascimbeni, A., et al., Dual task-related gait changes in patients with mild cognitive
impairment. Functional Neurology, 2015. 30(1): p. 59-65.
48. Taylor, M.E., et al., Gait parameter risk factors for falls under simple and dual task
conditions in cognitively impaired older people. Gait & Posture, 2013. 37(1): p. 126-30.
49. Wittwer, J.E., K.E. Webster, and K. Hill, The effects of a concurrent motor task on
walking in Alzheimer's disease. Gait & Posture, 2014. 39(1): p. 291-6.
50. Allali, G., et al., Impact of impaired executive function on gait stability. Dementia &
Geriatric Cognitive Disorders, 2008. 26(4): p. 364-9.
-52-
51. Allali, G., et al., Changes in gait while backward counting in demented older adults with
frontal lobe dysfunction. Gait & Posture, 2007. 26(4): p. 572-6.
52. Beauchet, O., et al., Test-retest reliability of stride time variability while dual tasking in
healthy and demented adults with frontotemporal degeneration. Journal of
NeuroEngineering and Rehabilitation, 2011. 8(1).
53. Lin, Y.C., et al., Comparison of motor performance of upper and lower extremities in
dual-task tests in patients with mild Alzheimer's dementia. Aging Clinical and
Experimental Research, 2016. 28(3): p. 491-496.
54. König, A., et al., Objective measurement of gait parameters in healthy and cognitively
impaired elderly using the dual-task paradigm. Aging Clinical & Experimental
Research, 2017. 29(6): p. 1181-1189.
55. Winblad, B., et al., Mild cognitive impairment – beyond controversies, towards a
consensus: report of the International Working Group on Mild Cognitive Impairment.
Journal of Internal Medicine, 2004. 256(3): p. 240-246.
56. McKhann, G., et al., Clinical diagnosis of Alzheimer's disease Report of the NINCDS‐
ADRDA Work Group* under the auspices of Department of Health and Human Services
Task Force on Alzheimer's Disease. Neurology, 1984. 34(7): p. 939-939.
57. Lezak, M.D., et al., Neuropsychological assessment. 2004: Oxford University Press,
USA.
58. D'Esposito, M., et al., Maintenance versus Manipulation of Information Held in Working
Memory: An Event-Related fMRI Study. Brain and Cognition, 1999. 41(1): p. 66-86.
59. St Clair-Thompson, H.L. and R.J. Allen, Are forward and backward recall the same? A
dual-task study of digit recall. Memory & Cognition, 2013. 41(4): p. 519-532.
60. Rabin, L.A., et al., Differential memory test sensitivity for diagnosing amnestic mild
cognitive impairment and predicting conversion to Alzheimer's disease.
Neuropsychology, development, and cognition. Section B, Aging, neuropsychology and
cognition, 2009. 16(3): p. 357-376.
-53-
61. Yesavage, J.A., et al., Development and validation of a geriatric depression screening
scale: a preliminary report. Journal of Psychiatric Research, 1982. 17(1): p. 37-49.
62. Bjelland, I., et al., The validity of the Hospital Anxiety and Depression Scale: an updated
literature review. Journal of Psychosomatic Research, 2002. 52(2): p. 69-77.
63. Hausdorff, J.M., et al., Dual-Task Decrements in Gait: Contributing Factors Among
Healthy Older Adults. Journals of Gerontology Series a-Biological Sciences and Medical
Sciences, 2008. 63(12): p. 1335-1343.
64. Hausdorff, J.M., J. Balash, and N. Giladi, Effects of Cognitive Challenge on Gait
Variability in Patients with Parkinson’s Disease. Journal of Geriatric Psychiatry and
Neurology, 2003. 16(1): p. 53-58.
65. Hamilton, F., et al., Walking and talking: an investigation of cognitive—motor dual
tasking in multiple sclerosis. Multiple Sclerosis Journal, 2009. 15(10): p. 1215-1227.
66. Beauchet, O., et al., Association of increased gait variability while dual tasking and
cognitive decline: results from a prospective longitudinal cohort pilot study.
GeroScience, 2017. 39(4): p. 439-445.
67. Beauchet, O., et al., Stops walking when talking: A predictor of falls in older adults?
European Journal of Neurology, 2009. 16(7): p. 786-795.
68. Herssens, N., et al., Do spatiotemporal parameters and gait variability differ across the
lifespan of healthy adults? A systematic review. Gait & Posture, 2018. 64: p. 181-190.
69. Beauchet, O., et al., Poor Gait Performance and Prediction of Dementia: Results From
a Meta-Analysis. Journal of the American Medical Directors Association, 2016. 17(6):
p. 482-490.
70. Pettersson, A., E. Olsson, and L.-O. Wahlund, Motor function in subjects with mild
cognitive impairment and early Alzheimer’s disease. Dementia and Geriatric Cognitive
Disorders, 2005. 19(5-6): p. 299-304.
-54-
71. Allali, G., et al., Gait phenotype from mild cognitive impairment to moderate dementia:
results from the GOOD initiative. European Journal of Neurology, 2016. 23(3): p. 527-
541.
72. Nasreddine, Z.S., et al., The Montreal Cognitive Assessment, MoCA: A Brief Screening
Tool For Mild Cognitive Impairment. Journal of the American Geriatrics Society, 2005.
53(4): p. 695-699.
73. Kelly, V.E., A.J. Eusterbrock, and A. Shumway-Cook, A Review of Dual-Task Walking
Deficits in People with Parkinson's Disease: Motor and Cognitive Contributions,
Mechanisms, and Clinical Implications. Parkinson's Disease, 2012. 2012(1): p. 14.
74. Rhea, C.K., et al., Noise and Complexity in Human Postural Control: Interpreting the
Different Estimations of Entropy. PLoS ONE [Electronic Resource], 2011. 6(3): p.
e17696.
75. Wittwer, J.E., K.E. Webster, and K. Hill, The effects of a concurrent motor task on
walking in Alzheimer's disease. Gait & Posture, 2014. 39(1): p. 291-296.
-55-
FIGURE LEGENDS
FIGURE 2.1 Flow diagram of the systematic review process.
FIGURE 2.2 Forest plot for within study comparison for single-task and dual-task for all gait
dynamic outcomes and all dual-task procedures
FIGURE 2.3 Forest plot for within study comparison for cognitively impaired and cognitively
healthy for all gait dynamic outcomes and all dual-task procedures
FIGURE 2.4 Forest plot for meta-analysis of within study comparison for MCI and dementia
for all gait dynamic outcomes and all dual-task conditions.
-56-
FIGURE 2.1 Flow diagram of the systematic review process
Scre
enin
g In
clud
ed
Elig
ibili
ty
Iden
tific
atio
n Additional records identified through other sources
(n = 5)
Search results (n = 16,524)
Records excluded bases on title or
abstract (n = 9,543)
Full-text articles excluded, with reasons
(n = 24) (n = 15 Not full text articles; n = 4 No gait variability measure; n = 1 No MCI/dementia diagnosis; n = 3 Complete data not extractable; n = 1 Same data set as an included study)
Studies included in qualitative synthesis
(n = 25)
Duplicates removed (n = 6,932)
Records identified through database searching
(n = 16,519)
Records after duplicates removed
(n = 9,592)
Full-text articles assessed for eligibility
(n = 49)
-57-
FIGURE 2.2 Forest plot for within study comparison for single-task and dual-task for all gait
dynamic outcomes and all dual-task procedures
-58-
Forest plot indicates that dual-task gait was more varied than single-task gait in cognitively
impaired adults, which reflects worse gait dynamics under dual-task conditions.
MCI=Mild Cognitive Impairment; SD=Standard deviation; Std=Standardized; CI=Confidence
interval; CV=Coefficient of variation; I2=Measures heterogeneity.
-59-
FIGURE 2.3 Forest plot for within study comparison for cognitively impaired and cognitively
healthy for all gait dynamic outcomes and all dual-task procedures
Forest plot indicates that dual-task cost was larger in cognitively impaired older adults than
cognitively healthy adults, which reflects worse gait dynamics.
MCI=Mild Cognitive Impairment; SD=Standard deviation; Std=Standardized; CI=Confidence
interval; CV=Coefficient of variation; I2=Measures heterogeneity.
-60-
FIGURE 2.4 Forest plot for meta-analysis of within study comparison for MCI and dementia for all gait dynamic outcomes and all dual-task
conditions
Forest plot indicates that dual-task cost was larger in dementia than MCI, which reflects worse gait dynamics.
MCI=Mild Cognitive Impairment; SD=Standard deviation; Std.=Standardized; CI=Confidence interval; CV=Coefficient of variation;
I2=Measures heterogeneity.
-61-
TABLE 2.1 Medline full electronic search strategy example
No. Searches
1 'dual task*' or 'dual-task*' or 'multi task*' or 'multi-task' or 'secondary task*' or
'attention task*' or 'cognitive task*' or 'motor task*' or 'two task*' or '2 task*').mp.
2 walk* or gait or locomot* or ambulat* or stride* or step* or 'double limb' or 'double-
limb' or 'double support' or 'swing time' or 'stride-to-stride' or 'stride to stride' or 'foot
clearance' or mobility or stability or instability or 'gait variability' or 'centre of
pressure' or 'center of pressure' or COP or 'centre of mass' or 'center of mass' or
COM or ataxia or McRoberts or 'Gait Up' or APDM or GAITRite or Axivity or AX3
or Opal* or Pedar or Zeno or Gyroscope* or Lyap* or fractal).mp.
3 1 and 2
No.=number; mp=title, abstract, original title, name of substance word, subject heading word,
floating sub-heading word, keyword heading word, protocol supplementary concept word, rare
disease supplementary concept word, unique identifier, synonyms.
-62-
TABLE 2.2 Cohort characteristics Author, Year
[reference]
Participants
Pathology
n Gender (%female)
Age
(Yr)
ST gait speed
(m/s)
MMSE
Sheridan,
2003
A:AD A:28 NR A:77.9±6.9 A:0.57±0.20 A:13.8±7.9
Camicioli,
2004
A:AD (Non-faller)
B:AD (Faller)
A:24
B:18
A:91.7
B:77.8
A:82.3±6.7
B:83.1±9.6
A:0.70±0.17
B:0.62±0.25
A:14.7±7.2
B:15.8±7.6
Camicioli,
2006
A:AD (w/≤3 EPS)
B:AD (w/>3 EPS)
A:13
B:29
A:100.0
B:79.3
A:80.5±6.5
B:83.6±8.5
A:0.79±0.15
B:0.62±0.18
A:14.4±7.1
B:15.5±7.5
Allali, 2007 A:Dementia A:16 A:83.6 A:87.5 NR A:22.1±3.6
Allali, 2008 C:Control
A:AD
B:CI (w/IEF)
C:22
A:16
B:18
C:91.0
A:69.0
B:83.0
C:79.5 (8)
A:78.5 (8)
B:79.5 (5)
NR C:30.0 (1)
A:22.0 (4)
B:20.5 (6)
Gillian, 2009 C:Control
A:MCI
B:AD
C:14
A:14
B:6
C:50.0
A:50.0
B:50.0
C:73.5
A:72.9
B:73.7
C:1.4±0.13
A:1.22±0.15
B:1.02±0.36
C:28.2±1.6
A:26.7±1.7
B:22.8±2.1
Allali, 2010 C:Control
A:Dementia (bvFTD)
B:AD
C:22
A:19
B:19
C:63.6
A:47.4
B:68.4
C:71.0±0.5
A:66.8±9.7
B:79.3±8.4
C:1.19±11.7
A:1.12±9.0
B:1.11±9.9
C:29.0±1.0
A:26.0±6.0
B:19.0±7.0
-63-
Beauchet,
2011
C:Control
A:Dementia (w/FTD)
C:69
A:14
C:43.5
A:7.1
C:75.5±4.3
A:65.7±9.8
NR A:23.3±6.6
Lamoth, 2011 C:Control
A:AD
C:13
A:13
C:53.8
A:69.2
C:79.4±5.6
A:82.6±4.3
C:0.95±0.21
A:0.88±0.27
C:28.2±1.1
A:18.0±3.5
Ijmker, 2012 C:Control (older)
A:D
C:14
A:15
C:14.3
A:13.3
C:76.9±4.1
A:81.7±6.3
C:1.14±0.11
A:0.67±2.1
C:28.5±1.2
A:19.6±3.6
Montero-
Odasso, 2012
C:Control
A:MCI
C:25
A:43
C:88.0
A:54.0
C:71.5±4.1
A:75.1±6.3
NR C:29.5±0.6
A:27.8±1.6
Muir, 2012 C:Control
A:MCI
B:AD
C:22
A:29
B:23
C:88.0
A:59.0
B:61.0
C:71.0±5.0
A:73.6±6.2
B:77.5±5.0
C:1.36±0.24
A:1.16±0.21
B:1.11±0.14
C:29.5±0.6
A:27.5±1.9
B:24.2±2.3
Taylor, 2013 A:CI (Non-multiple
fallers)
B:CI (Multiple fallers)
A:41
B:22
A:46.3
B:45.5
A:80.7±6.8
B:82.5±6.9
A:0.94±0.24
B:0.79±0.30
A:24.8±3.6
B:22.7±5.1
Beauchet,
2014
A:AD A:86 A:68.6 A:82.6±5.5 A:0.63±0.21 A:17.6±5.5
Hsu, 2014 C:Control
A:AD
C:50
A:21
C:40.0
A:52.4
C:59.9±4.6
A:61.5±4.9
C:1.38±0.17
A:1.25±0.16
C:28.4±1.6
A:23.0±3.2
-64-
Montero-
Odasso, 2014
A:MCI (Na)
B:MCI (a)
A:22
B:42
A:63.6
B:42.9
A:74.2±6.5
B:77.3±7.3
A:1.09±0.19
B:1.00±0.22
A:29.1±0.8
B:27.2±2.1
Wittwer,
2014
A:AD A:30 A:50.0 A:80.2±5.8 A:1.12±0.27 A:20.6±5.1
Nascimbeni,
2015
C:Control
A:MCI
C:10
A:13
C:40.0
A:15.4
C:72.0±3.9
A:76.0±3.9
C:0.97±0.15
A:0.83±0.21
NR
Gillian, 2016 A:MCI (future AD)
B:MCI
A:9
B:4
A:44.4
B:50.0
A:74.4±4.16
B:70.0±2.16
A:1.15±0.13
B:1.29±0.10
A:26.1±1.5
B:27.3±1.7
Lin, 2016 C:Control
A:AD
C:10
A:10
C:80.0
A:80.0
C:73.8±6.1
A:74.0±8.6
NR C:29.4±0.7
A:17.7±4.1
Martinez-
Ramirez,
2016
C:Control (frail)
A:MCI (frail)
C:20
A:11
C:70.0
A:72.7
C:93.4±3.2
A:92.4±4.2
C:0.68±0.26
A:0.77±0.18
NR
Auvinet, 2017 A:MCI A:24 A:33.3 A:76.4±5.8 A:1.0±0.3 A:26.2±2.1
Gschwind,
2017
A:MCI A:50 A:50.0 A:68.5±8.4 A:1.27±0.18 NR
Konig, 2017 C:Control
A:AD
B:MCI
C:22
A:23
B:24
C:68.2
A:47.8
B:66.7
C:73.0±7.0
A:77.0±9.0
B:75.0±9.0
NR C:28.4±1.5
A:17.0±4.6
B:24.8±3.2
-65-
Lee, 2018 C:Control
A:MCI
C:8
A:8
NR C:66.1±1.6
A:66.5±1.9
C:1.06±0.06
A:1.00±0.18
C:28.1±0.8
A:21.0±0.8
Results are presented as Mean±Standard Deviation or Median (interquartile range); n=number;
%=percent; Yr=years; kg=kilograms; m/s=meters per second; C=control; A/B=intervention
group A or B; w/=with; IEF=impaired executive function; D=dementia; AD=Alzheimer's
disease; FTD=Frontotemporal degeneration; EPS=extra-pyramidal signs; MCI=mild cognitive
impairment; CI=cognitive impairment; bvFTD=behavioral variant of frontotemporal
degeneration; NR=not reported; ST=Single-task.
-66-
TABLE 2.3 Risk of bias assessment Author, Year [reference]
Reporting, item no. External validity, item no.
Internal validity, item no. Total /24 Bias Confounding
1 2 3 4 5 6 7 8 10 11 12 13 14 15 16 18 19 20 21 22 23 24 25 27 Sheridan, 2003 1 1 1 1 0 1 1 0 1 0 0 1 0 0 1 1 1 1 1 0 0 0 0 0 13 Camicioli, 2004 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 0 0 1 0 14 Camicioli, 2006 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 0 0 1 0 14 Allali, 2007 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 1 0 0 0 12 Allali, 2008 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 1 0 1 0 13 Gillian, 2009 1 1 1 1 1 1 1 0 0 0 0 1 0 0 1 1 1 1 1 0 0 0 0 0 13 Allali, 2010 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 1 0 1 0 15 Beauchet, 2011 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 11 Lamoth, 2011 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 1 13 Ijmker, 2012 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 0 0 0 1 12 Montero-Odasso, 2012 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 1 0 1 1 16 Muir, 2012 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 1 0 0 0 14 Taylor, 2013 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 12 Beauchet, 2014 1 1 1 1 1 1 1 0 1 1 0 1 0 0 1 1 1 1 1 1 0 0 0 1 17 Hsu, 2014 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 12 Montero-Odasso, 2014 1 1 1 1 1 1 1 0 1 0 0 0 0 0 1 1 1 1 1 0 1 0 1 1 16 Wittwer, 2014 1 1 1 1 0 1 1 1 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 13 Nascimbeni, 2015 1 1 1 1 1 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 1 0 1 1 15 Gillian, 2016 1 1 1 1 0 1 1 0 1 0 0 1 0 0 1 1 1 1 1 1 0 0 0 0 14 Lin, 2016 1 0 0 1 1 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 1 0 0 0 11 Martinez-Ramirez, 2016 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 0 0 0 1 12 Auvinet, 2017 1 1 1 1 1 1 1 0 1 1 0 1 0 0 1 1 1 1 1 0 0 0 1 0 16 Gschwind, 2017 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1 1 1 1 1 1 1 0 0 1 17
-67-
Konig, 2017 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 1 0 0 0 1 0 15 Lee, 2018 1 1 0 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 10
Results are presented for each criterion with a total score presented under the ‘Total’ column with a maximum score of 24. The criteria that were
not included for scoring were: items 9, 17 and 26 (relating to follow up). Items modified to be consistent with the scoring procedure included:
item number 5 (relating to confounder distribution), and item number 27 (relating to statistical power), i.e., alteration from a 0 to 2 scale, and a 0
to 5 scale, to a 'no, 0; unable to determine, 0; and yes, 1'. Criterion 23 was altered for increased specificity to read: ‘Was the order of the walking
tasks (single and dual) randomized for study subjects?’.
-68-
TABLE 2.4 Dual-task procedure characteristics
Author, Year [reference]
Measurement system
Dual-task type Dual -task condition(s)
Measured distance walked (m)
Number of trials per condition (n)
Order of walking randomized
Sheridan, 2003
Footswitch sensors
Working memory Forward digit span 15.24-152.4 (up to 10 laps) Avg distance 106.68 (7laps)
NR Set order
Camicioli, 2004
GAITRite Working memory Forward counting from 1 by 1s
3.66 ST: 2 (used 2nd trial data only), DT: 1
Set order
Camicioli, 2006
GAITRite Working memory Forward counting from 1 by 1s
3.66 ST: 2 DT: 1
Set order
Allali, 2007 GAITRite Working memory Mental tracking
Forwards counting Backwards counting
10 1 Randomized
Allali, 2008 GAITRite Mental tracking Backwards counting from 50 by 1s
10 NR Randomized
Gillian, 2009 Locometrix Mental tracking Backwards counting from 50 by 1s
30 3 (Used 2nd trial data only)
NR
Allali, 2010 SMTEC footswitch
Mental tracking Backwards counting from 50 by 1s
10 1 Randomized
Beauchet, 2011
GAITRite & SMTEC footswitch system
Mental tracking Backwards counting from 50 by 1s
3.5 (GAITRite) & 10 (SMTEC)
2 NR
-69-
Lamoth, 2011 DynaPort MiniMod
Verbal fluency Name words starting with either 'R' or 'G'
~160m (3 min in 40m corridor)
1 NR
Ijmker, 2012 DynaPort MiniMod
Verbal fluency Name animals, occupations or words starting with either 'R', 'G' or 'P' in 1 minute.
10 1 NR
Montero-Odasso, 2012
GAITRite Mental tracking Verbal fluency
Subtracting serial 7s from 100 Naming animals
6 NR (1st trial as a practice)
Randomized
Muir, 2012 GAITRite Mental tracking Verbal fluency Mental tracking
Backwards counting from 100 by 1s Naming animals Subtracting serial 7s from 100
6 NR (1st trial as a practice)
Randomized
Taylor, 2013 GAITRite Motor Mental tracking
Carrying a glass filled (10mm from rim) of water Backwards counting from 30 by 1s
4.6 2 NR
Beauchet, 2014
GAITRite Mental tracking Backwards counting from 50 by 1s
7.92 1 NR
Hsu, 2014 Wearable device Mental tracking Backwards counting from 100 by 1s
40 NR NR
-70-
Montero-Odasso, 2014
GAITRite Mental tracking Mental tracking Verbal fluency
Backwards counting from 100 by 1s Subtracting serial 7s from 100 Naming animals
6 NR (1st trial as a practice)
Randomized
Wittwer, 2014 GAITRite Motor Carrying a tray with two empty glasses using both hands
8.3 2 to 4 (mean of trials used)
Set order
Nascimbeni, 2015
STEP 32 system & 3 footswitch sensors
Verbal fluency Verbal memory Mental tracking
Naming words beginning with F, A or S for 1 minute Short story recall Backwards counting from either 378 or 283 by 1s
12 1 Randomized
Gillian, 2016 Locometrix Mental tracking Backwards counting from 50 by 1s
30 1 Set order
Lin, 2016 Vicon MX infrared camera and 3 AMTI force plates
Working memory Mental tracking
Forward 3-digit working task Backwards 3-digit working task
8 6 (1st trial as a practice)
Randomized
Martinez-Ramirez, 2016
Orientation Tracker MTx
Mental tracking Verbal fluency
Backwards counting from 100 by 1s Naming animals
5 NR NR
-71-
Auvinet, 2017 Locometrix Mental tracking Backwards counting from 50 by 1s
30 1 Set order
Gschwind, 2017
GAITRite Mental tracking Verbal fluency
Backwards counting from 50 by 2s Naming animals
10 1 Randomized
Konig, 2017 CE-marked accelerometer
Mental tracking Backwards counting from 305 by 1s.
~20 (10 up and 10 back plus turn)
1 NR
Lee, 2018 3D movement analysis
Mental tracking Backwards counting from 100 by 1s
12 1 Set order
m=meters; n=number; ST=Single-task; DT=Dual-task; NR=not reported; all verbal fluency tasks that used a word naming task using a specific
letter stated that the letter choice was predetermined; further detail for the task in Sheridan, 2003 was to repeat a list of random single-digit numbers
forward starting with 2 digits and progressing to 8 digits.
-72-
TABLE 2.5a Gait outcomes for Mild Cognitive Impairment group: single-task vs. dual-task comparisons Author, Year [reference]
Pathology Task Single-task Mean±SD
Dual-task Mean±SD
Mean Difference (95% CI)
Between group ES (95% CI)
Stride time CV
Montero-Odasso, 2012
MCI Backwards counting from 100 by 7s
2.68±1.31 9.84±10.13 7.16 [6.38, 7.94] 5.37 [4.03, 6.70]
MCI Animal naming 2.68±1.31 7.16±7.76 4.48 [3.70, 5.26] 3.36 [2.40, 4.31]
Muir, 2012 MCI Backwards counting from 100 by 1s
2.59±1.47 4.06±2.37 1.47 [0.40, 2.54] 0.97 [0.20, 1.75]
MCI Animal naming 2.59±1.47 8.02±8.88 5.43 [4.36, 6.50] 3.59 [2.36, 4.82]
MCI Backwards counting
from 100 by 7s 2.59±1.47 10.07±9.29 7.48 [6.41, 8.55] 4.95 [3.39, 6.50]
Montero-Odasso, 14
MCI (na) Backwards counting from 100 by 1s
2.40±1.38 2.90±0.98 0.50 [-0.65, 1.65] 0.35 [-0.49, 1.19]
MCI (a) Backwards counting from 100 by 1s
3.33±2.60 4.81±3.73 1.51 [-0.06, 3.08] 0.57 [-0.05, 1.19]
MCI (na) Backwards counting from 100 by 7s
2.40±1.38 3.82±2.10 1.42 [0.27, 2.57] 0.99 [0.09, 1.89]
MCI (a) Backwards counting from 100 by 7s
3.33±2.60 5.63±5.00 2.30 [0.73, 3.87] 0.87 [0.23, 1.50]
-73-
MCI (na) Animal naming 2.40±1.38 4.83±3.53 2.43 [1.28, 3.58] 1.69 [0.69, 2.70]
MCI (a) Animal naming 3.33±2.60 6.47±5.71 3.14 [1.57, 4.71] 1.18 [0.52, 1.85]
Nascimbeni, 2015 MCI Word naming with 'F', 'A', or 'S'
3.17±1.12 5.52±3.04 2.35 [1.13, 3.57] 1.95 [0.54, 3.36]
MCI Short story recall 3.17±1.12 5.42±1.81 2.25 [1.03, 3.47] 1.87 [0.48, 3.26]
MCI Backwards counting
from 378 or 283 by 1s 3.17±1.12 5.07±3.30 1.90 [0.68, 3.12] 1.58 [0.27, 2.89]
Gschwind, 2017
MCI Backwards counting from 50 by 12s
1.95±0.90 6.40±16.20 4.45 [3.95, 4.95] 4.87 [3.73, 6.00]
MCI Animal naming 1.95±0.90 8.65±16.10 6.70 [6.20, 7.20] 7.33 [5.73, 8.92]
Lee, 2018 MCI Backwards counting from 100 by 1s
2.44±0.99 11.21±6.84 8.77 [7.40, 10.14] 7.70 [2.23, 13.18]
Step regularity
Gillian, 2009 MCI Backwards counting from 50 by 1s
287.00±29.00 224.00±47.00 -63.00 [-93.38, -32.62] 2.03 [0.66, 3.41]*
Martinez-Ramirez, 2016
MCI (frail) Backwards counting from 100 by 1s
0.58±0.18 0.51±0.21 -0.07 [-0.28, 0.14] 0.36 [-0. 85, 1.56]*
MCI (frail) Naming animals 0.58±0.18 0.37±0.21 -0.21 [-0.42, 0.00] 1.07 [-0.24, 2.38]*
-74-
Stride regularity
Gillian, 2016
MCI (future AD) Backwards counting from 50 by 1s
286.20±37.45 220.67±254.88 -65.53 [-114.77, -16.29]
1.56 [-0.07, 3.18]*
MCI (-) Backwards counting from 50 by 1s
298.00±22.46 254.88±32.86 -43.12 [-87.14, 0.09] 1.10 [-5.41, 5.61]*
Martinez-Ramirez, 2016
MCI (frail) Backwards counting from 100 by 1s
0.59±0.16 0.44±0.19 -0.15 [-0.34, 0.04] 0.86 [-0.41, 2.13]*
MCI (frail) Naming animals 0.59±0.16 0.38±0.14 -0.21 [-0.40, -0.02] 1.20 [-0.14, 2.54]*
Auvinet, 2017 MCI Backwards counting from 50 by 1s
214.70±54.20 156.60±65.20 -58.10 [-14.73, -101.47]
1.03 [0.17, 1.90]*
Step time CV
Martinez-Ramirez, 2016
MCI (frail) Backwards counting from 100 by 1s
0.09±0.02 0.11±0.05 0.02 [0.00, 0.04] 0.91 [-0.36, 2.19]
MCI (frail) Naming animals 0.09±0.02 0.15±0.06 0.06 [0.04, 0.08] 2.74 [0.88, 4.60]
Step time variance
Konig, 2017 MCI Backwards counting from 305 by 1s
5.70±4.50 6.80±5.30 1.10 [-2.50, 4.70] 0.24 [-0.57, 1.04]
Approximate entropy
Martinez-Ramirez, 2016
MCI (frail) Backwards counting from 100 by 1s
0.16±0.12 2.12±1.65 1.96 [1.82, 2.10] 14.93 [7.05, 22.81]
MCI (frail) Naming animals 0.16±0.12 3.7±2.41 3.54 [3.40, 3.68] 26.97 [12.85,
41.09]
-75-
SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; MCI=mild cognitive impairment; (a)=amnestic;
(na)=non-amnestic; *=ES was reversed to indicate change in the direction of other outcome measures. ESs were calculated as standardized mean
difference and 95% CI.
-76-
TABLE 2.5b Gait outcomes for dementia (including Alzheimer’s disease) group: single-task vs. dual-task comparisons
Author, Year [reference]
Pathology Task Single-task Mean±SD
Dual-task Mean±SD
Mean Difference (95% CI)
Between group ES (95% CI)
Stride time CV Sheridan, 2003 AD Forward digit
span 8.50±3.40 11.10±5.50 2.60 [0.08, 5.12] 0.74 [-0.03, 1.51]
Camicioli, 2006
AD (w/≤3 EPS)
Forwards counting from 1 by 1s
4.04±1.94 4.65±2.83 0.61 [-1.51, 2.71] 0.29 [-0.81, 1.39]
AD (w/>3EPS) Forwards
counting from 1 by 1s
4.36±3.81 7.78±12.20 3.42 [0.65, 6.19] 0.87 [0.11, 1.64]
Allali, 2007 Cognitive impairment
Forwards counting
4.00±2.20 7.60±10.00 3.60 [1.44, 5.76] 1.55 [0.39, 2.71]
Cognitive impairment
Backwards counting
4.00±2.20 15.40±16.10 11.40 [9.24, 13.56] 4.90 [2.71, 7.15]
Allali, 2008
AD Backwards counting from 50 by 1s
1.17±0.56 4.10±3.47 2.93 [2.38, 3.48] 4.95 [2.74, 7.15]
Cognitive impairment (w/IEF)
Backwards counting from 50 by 1s
2.77±2.02 14.33±13.72 11.56 [9.69, 13.43] 5.45 [3.23, 7.67]
-77-
Allali, 2010
bvFTD Backwards counting from 50 by 1s
7.70±8.20 8.30±6.20 0.60 [-6.78, 7.98] 0.07 [-0.83, 0.97]
AD Backwards counting from 50 by 1s
3.10±1.20 6.00±3.10 2.90 [1.82, 3.98] 2.31 [1.09, 3.53]
Beauchet, 2011 Dementia (w/IEF)
Backwards counting from 50 by 1s
5.00±2.50 7.60±6.70 2.60 [-0.02, 5.22] 0.97 [-0.16, 2.10]
Lamoth, 2011 AD Word naming with 'R' or 'G'
4.20±2.7 3.67±1.67 -0.53 [-3.47, 2.41] -0.18 [-1.28, 0.91]
Ijmker, 2012 AD Word naming with 'R', 'G' or 'P'
9.88±5.28 12.88±6.78 3.00 [-2.36, 8.36] 0.53 [-0.50, 1.57]
Muir, 2012 AD Backwards counting from 100 by 1s
2.67±1.08 4.86±2.74 2.19 [1.31, 3.07] 1.95 [0.93, 2.98]
AD Animal naming 2.67±1.08 9.04±8.94 6.37 [5.49, 7.25] 5.68 [3.70, 7.67]
AD Backwards
counting from 100 by 7s
2.67±1.08 12.49±12.33 9.82 [8.94, 10.70] 8.76 [5.86, 11.66]
Beauchet, 2014
AD Backwards counting from 50 by 1s
5.61±4.00 8.85±7.40 3.24 [1.55, 4.93] 0.80 [0.36, 1.24]
Hsu, 2014 AD Backwards counting from 100 by 1s
2.31±0.66 31.31±26.83 29.00 [28.43, 29.57] 42.18 [28.00, 56.36]
-78-
Wittwer, 2014 AD Carrying a tray with two empty glasses using both hands
2.40±0.80 2.80±0.80 0.40 [-0.17, 0.97] 0.49 [-0.24, 1.21]
Lin, 2016 AD Forwards 3-digit span
5.20±1.90 5.80±5.00 0.60 [-1.76, 2.96] 0.29 [-0.96, 1.54]
AD Backwards 3-
digit span 5.20±1.90 9.90±3.80 4.70 [2.34, 7.06] 2.23 [0.47, 4.00]
Stride length CV
Camicioli, 2006
AD (w/≤3 EPS)
Forwards counting from 1 by 1s
5.73±3.36 4.27±2.33 -1.46 [-2.34, 5.26] -0.40 [-0.75, 1.55]
AD (w/>3 EPS)
Forwards counting from 1 by 1s
5.78±2.25 5.59±2.51 -0.19 [-1.48, 1.86] -0.08 [-0.66, 0.82]
Taylor, 2013
Cognitive impairment (non-multiple faller)
Carrying a glass filled with water
2.63±1.62 3.17±2.05 0.54 [-0.46, 1.54] 0.33 [-0.30, 0.95]
Cognitive impairment (multiple faller)
Carrying a glass filled with water
4.20±2.54 5.76±4.66 1.56 [-0.56, 3.68] 0.59 [-0.27, 1.45]
Cognitive impairment (non-multiple faller)
Backwards counting from 30 by 1s
2.63±1.62 3.74±2.83 1.11 [0.11, 2.11] 0.67 [0.03, 1.31]
-79-
Cognitive impairment (multiple faller)
Backwards counting from 30 by 1s
4.20±2.54 6.34±5.03 2.14 [0.02, 4.26] 0.81 [-0.07, 1.69]
Wittwer, 2014 AD Carrying a tray with two empty glasses using both hands
3.20±1.00 3.90±1.50 0.70 [-0.02, 1.42] 0.68 [-0.06, 1.42]
Lin, 2016 AD Forwards 3-digit span
6.70±5.30 7.70±3.40 1.00 [-5.57, 7.57] 0.17 [-1.07, 1.41]
AD Backwards 3-
digit span 6.70±5.30 10.40±2.30 3.70 [-2.87, 10.27] 0.63 [-0.66, 1.92]
Swing time CV
Camicioli, 2006
AD (w/≤3 EPS)
Forwards counting from 1 by 1s
6.15±4.19 8.58±4.53 2.43 [-2.14, 7.00] 0.54 [-0.58, 1.66]
AD (w/>3 EPS)
Forwards counting from 1 by 1s
7.95±3.92 11.09±7.40 3.14 [0.28, 6.00] 0.78 [0.02, 1.54]
Taylor, 2013
Cognitive impairment (non-multiple faller)
Carrying a glass filled with water
6.25±3.39 6.74±5.30 0.49 [-1.59, 2.57] 0.14 [-0.47, 0.75]
Cognitive impairment (multiple faller)
Carrying a glass filled with water
8.07±4.37 10.65±7.67 2.58 [-1.07, 6.23] 0.57 [-0.29, 1.42]
-80-
Cognitive impairment (non-multiple faller)
Backwards counting from 30 by 1s
6.25±3.39 9.42±7.31 3.17 [1.09, 5.25] 0.92 [0.27, 1.56]
Cognitive impairment (multiple faller)
Backwards counting from 30 by 1s
8.07±4.37 15.06±11.25 6.99 [3.34, 10.64] 1.54 [0.56, 2.51]
Hsu, 2014 AD Backwards counting from 100 by 1s
2.67±0.69 17.24±12.99 14.57 [13.98, 15.16] 20.27 [13.42, 27.13]
Walking speed CV
Beauchet, 2014
AD Backwards counting from 50 by 1s
8.25±5.00 11.49±7.10 3.24 [1.13, 5.35] 0.64 [0.21, 1.08]
Base of support CV
Camicioli, 2006
AD (w/≤3 EPS)
Forwards counting from 1 by 1s
23.7±13.45 29.6±22.80 5.90 [-8.77, 20.57] 0.41 [-0.70, 1.54]
AD (w/>3 EPS)
Forwards counting from 1 by 1s
18.7±15.3 19.3±11.90 0.60 [-10.54, 11.74] 0.04 [-0.69, 0.77]
Double support time CV
Camicioli, 2006
AD (w/≤3 EPS)
Forwards counting from 1 by 1s
11.20±6.96 8.65±4.88 -2.55 [-10.14, 5.04] -0.34 [-0.44, 0.76]
-81-
AD (w/>3 EPS)
Forwards counting from 1 by 1s
15.10±11.60 19.70±20.3 4.60 [-3.85, 13.05] 0.39 [-0.35, 1.12]
Left stride length CV
Camicioli, 2004
AD (non-faller) Forwards counting from 1 by 1s
5.79±2.95 5.25±3.47 -0.54 [-2.90, 1.82] -0.18 [-0.98, 0.63]
AD (faller) Forwards counting from 1 by 1s
5.89±3.20 5.67±3.48 -0.22 [-3.18, 2.74] -0.07 [-0.99, 0.86]
Left base of support CV
Camicioli, 2004
AD (non-faller) Forwards counting from 1 by 1s
15.13±10.82 24.71±21.95 9.58 [0.92, 18.24] 0.85 [0.01, 1.70]
AD (faller) Forwards counting from 1 by 1s
22.22±21.58 18.06±10.11 -4.16 [-24.10, 15.78] -0.18 [-1.11, 0.74]
Right stride length CV
Camicioli, 2004
AD (non-faller) Forwards counting from 1 by 1s
5.38±2.67 5.25±3.82 -0.13 [-2.27, 2.01] -0.05 [-0.85, 0.75]
AD (faller) Forwards counting from 1 by 1s
6.11±3.60 4.83±2.57 -1.28 [-4.61, 2.05] -0.34 [-1.27, 0.59]
Right base of support CV
Camicioli, 2004
AD (non-faller) Forwards counting from 1 by 1s
22.67±21.39 23.42±22.61 0.75 [-16.37, 17.87] 0.03 [-0.77, 0.83]
-82-
AD (faller) Forwards counting from 1 by 1s
22.06±21.16 22.72±21.61 0.66 [-18.89, 20.21] 0.03 [-0.89, 0.95]
Step regularity
Gillian, 2009 AD Backwards counting from 50 by 1s
227.00±82.00 139.00±81.00 -88.00 [-219.22, -43.22]
0.86 [-0.94, 2.66]*
Stance time CV
Hsu, 2014 AD Backwards counting from 100 by 1s
3.31±0.84 43.31±37.32 40.00 [39.28, 40.72] 45.71 [30.35, 61.80]
Stance period CV
Hsu, 2014 AD Backwards counting from 100 by 1s
4.78±5.48 12.3±10.18 7.52 [2.83, 12.21] 1.32 [0.35, 2.32]
Swing period CV
Hsu, 2014 AD Backwards counting from 100 by 1s
5.74±7.80 18.95±15.09 13.21 [6.53, 19.89] 1.63 [0.61, 2.64]
Step time variance
Konig, 2017 AD Backwards counting from 305 by 1s
6.70±7.10 10.20±9.90 3.50 [-2.31, 9.31] 0.47 [-0.36, 1.31]
Stride time DFA Lamoth, 2011 AD Word naming
with 'R' or 'G' 0.84±0.16 0.84±0.11 0.00 [-0.17, 0.17] 0.00 [-1.09, 1.09]
-83-
SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; AD=Alzheimer’s disease; w/=with; EPS=extra-
pyramidal signs; IEF=Impaired executive function; bvFTD=Behvioural variant frontotemporal degeneration; FTD=Frontotemporal degeneration;
*=ES was reversed to indicate change in the direction of other outcome measures. ESs were calculated as standardized mean difference and 95%
CI.
-84-
TABLE 2.6a Gait outcomes for cognitive status comparisons: Control vs. Mild Cognitive Impairment
Author, Year [reference] Task
Control group Cognitive impairment group
Mean Difference (95% CI)
Between group ES (95% CI) Single-task
Mean±SD Dual-task Mean±SD
Single-task Mean±SD
Dual-task Mean±SD
Stride time CV
Montero-Odasso, 2012
Backwards counting from 100 by 7s
1.86±0.66 3.74±3.31 2.68±1.31 9.84±10.13 5.28 [4.73, 5.83] 4.67 [3.72,5.61]
Naming animals 1.86±0.66 3.59±2.95 2.68±1.31 7.16±7.76 2.75 [2.20, 3.30] 2.43 [1.78, 3.08]
Muir, 2012 Backwards counting from 100 by 1s
1.72±0.66 2.12±1.35 2.59±1.47 4.06±2.37 1.07 [0.41, 1.73] 0.88 [0.30, 1.47]
Naming animals 1.72±0.66 2.69±1.57 2.59±1.47 8.02±8.88 4.46 [3.80, 5.12] 3.68 [2.76, 4.61]
Backwards counting from 100 by 7s
1.72±0.66 3.14±2.18 2.59±1.47 10.07±9.29 6.06 [5.40, 6.72] 5.00 [3.85, 6.16]
Nascimbeni, 2015
Word naming with 'F', 'A', or 'S'
3.58±1.99 4.44±1.69 3.17±1.12 5.52±3.04 1.49 [0.21, 2.77] 0.92 [0.05, 1.80]
-85-
Short story recall 3.58±1.99 5.57±2.60 3.17±1.12 5.42±1.81 0.26 [-1.02, 1.54] 0.16 [-0.66, 0.99]
Backwards counting from 378 or 283 by 1s
3.58±1.99 3.93±1.70 3.17±1.12 5.07±3.30 2.09 [0.81, 3.37] 1.30 [0.38, 2.22]
Lee, 2018 Backwards counting from 100 by 1s
2.77±1.05 4.49±1.90 2.44±0.99 11.21±6.84 7.05 [6.05, 8.05] 6.53 [3.75, 9.32]
Step regularity Gillian, 2009 Backwards counting
from 50 by 1s 276.00±35.00 258.00±38.00 287.00±29.00 224.00±47.00 -45.00 [-68.81, -21.19] 1.36 [0.52, 2.19]*
Martinez-Ramirez, 2016
Backwards counting from 100 by 1s
0.48±0.21 0.47±0.17 0.58±0.18 0.51±0.21 -0.06 [-0.21, 0.09] 0.29 [0.45, 1.03]*
Naming animals 0.48±0.21 0.40±0.21 0.58±0.18 0.37±0.21 -0.13 [-0.28, 0.02] 0.63 [0.12, 1.39]*
Stride regularity Martinez-Ramirez, 2016
Backwards counting from 100 by 1s
0.45±0.20 0.43±0.20 0.59±0.16 0.44±0.19 -0.13 [-0.27, 0.01] 0.68 [0.08, 1.43]*
Naming animals 0.45±0.20 0.36±0.22 0.59±0.16 0.38±0.14 -0.12 [-0.26, 0.02] 0.62 [0.13,1.38]*
Step time CV
-86-
Martinez-Ramirez, 2016
Backwards counting from 100 by 1s
0.12±0.05 0.13±0.06 0.09±0.02 0.11±0.05 0.01 [-0.02, 0.04] 0.23 [-0.51, 0.97]
Naming animals 0.12±0.05 0.14±0.07 0.09±0.02 0.15±0.06 0.04 [0.01, 0.07] 0.92 [0.15, 1.70]
Step time variance Konig, 2017 Backwards counting
from 305 by 1s 4.50±4.90 3.90±5.40 5.70±4.50 6.80±5.30 1.70 [-1.02, 4.42] 0.36 [-0.23, 0.94]
Approximate entropy
Martinez-Ramirez, 2016
Backwards counting from 100 by 1s
0.27±0.36 2.42±2.89 0.16±0.12 2.12±1.65 -0.19 [-0.41, 0.03] -0.62 [-1.37, 0.14]
Naming animals 0.27±0.36 3.67±4.12 0.16±0.12 3.7±2.41 0.14 [-0.08, 0.36] 0.45 [-0.29, 1.20]
SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; *=ES was reversed to indicate change in the
direction of other outcome measures. ESs were calculated as standardized mean difference and 95% CI.
-87-
TABLE 2.6b Gait outcomes for cognitive status comparisons: Control vs. dementia (including Alzheimer’s disease)
Author, Year [reference] Task
Control group Cognitive impairment group
Mean Difference (95% CI)
Between group ES (95% CI) Single-task
Mean±SD Dual-task Mean±SD
Single-task
Mean±SD
Dual-task Mean±SD
Stride time CV
Allali, 2008 Backwards counting from 50 by 1s
1.47±1.11 2.17±1.43 1.17±0.56 4.10±3.47 2.23 [1.64, 2.82] 2.37 [1.51, 3.22]
Backwards counting from 50 by 1s
1.47±1.11 2.17±1.43 2.77±2.02 14.33±13.72 10.86 [9.87, 11.85] 6.74 [5.06, 8.41]
Allali, 2010 Backwards counting from 50 by 1s
1.70±0.50 2.70±0.90 7.70±8.00 8.30±6.20 -0.40 [-3.83, 3.03] -0.07 [-0.68, 0.54]
Backwards counting from 50 by 1s
1.70±0.50 2.70±0.90 3.10±1.20 6.00±3.10 1.90 [1.35, 2.45] 2.09 [1.32, 2.87]
Beauchet, 2011 Backwards counting from 50 by 1s
1.30±1.00 1.70±1.40 5.00±2.50 7.60±6.70 2.20 [1.42, 2.98] 1.61 [0.98, 2.23]
Lamoth, 2011 Word naming with 'R' or 'G'
2.95±1.77 5.00±2.67 4.2±2.70 3.67±1.67 -2.58 [-4.33, -0.83] -1.09 [-1.93, -0.26]
Ijmker, 2012 Word naming with 'R', 'G' or 'P'
3.51±0.88 4.26±1.0 9.88±5.28 12.88±6.78 2.25 [-0.55, 5.05] 0.57 [-0.18, 1.31]
-88-
Muir, 2012 Backwards counting from 100 by 1s
1.72±0.66 2.12±1.35 2.67±1.08 4.86±2.74 1.79 [1.26, 2.32] 1.95 [1.23, 2.68]
Animal naming 1.72±0.66 2.69±1.57 2.67±1.08 9.04±8.94 5.40 [4.87, 5.93] 5.90 [4.49, 7.30]
Backwards counting from 100 by 7s
1.72±0.66 3.14±2.18 2.67±1.08 12.49±12.33 8.40 [7.87, 8.93] 9.17 [7.10, 11.24]
Hsu, 2014 Backwards counting from 100 by 1s
2.02±0.78 14.38±18.37
2.31±0.66 31.31±26.83 16.64 [16.26, 17.02] 22.03 [18.26, 25.79]
Lin, 2016 Forwards 3-digit span
3.40±1.90 4.30±1.20 5.20±1.90 5.80±5.00 -0.30 [-1.97, 1.37] -0.15 [-1.03, 0.73]
Backwards 3-digit span
3.40±1.90 4.20±1.80 5.20±1.90 9.90±3.80 3.90 [2.23, 5.57] 1.97 [0.86, 3.08]
Stride length CV Lin, 2016 Forwards 3-
digit span 3.40±1.00 4.20±1.70 6.70±5.30 7.70±3.40 0.20 [-3.14, 3.54] 0.05 [-0.83, 0.93]
Backwards 3-digit span
3.40±1.00 7.50±13.00
6.70±5.30 10.40±2.30 -0.40 [-3.74, 2.94] -0.10 [-0.98, 0.78]
Step regularity Gillian, 2009 Backwards
counting from 50 by 1s
276.00±35.00
258.00±38.00
227.00±82.00
139.00±81.00
-70.00 [-120.17, -19.83]
1.28 [0.22, 2.33]*
-89-
Stance time CV
Hsu, 2014 Backwards counting from 100 by 1s
3.13±1.07 19.47±27.4
3.31±0.84 43.31±37.32 23.66 [23.15, 24.17] 23.20 [19.24, 27.16]
Swing time CV Hsu, 2014 Backwards
counting from 100 by 1s
2.47±0.64 13.23±8.77
2.67±0.69 17.24±12.99 6.81 [6.48, 7.14] 10.29 [8.47, 12.10]
Stance period CV Hsu, 2014 Backwards
counting from 100 by 1s
1.8±0.47 5.97±4.08 4.78±5.48 12.3±10.18 3.35 [1.83, 4.87] 1.11 [0.57, 1.66]
Swing period CV Hsu, 2014 Backwards
counting from 100 by 1s
2.17±0.58 6.88±5.91 5.74±7.8 18.95±15.09 8.50 [6.35, 10.65] 1.99 [1.38, 2.60]
Step time variance Konig, 2017 Backwards
counting from 305 by 1s
4.50±4.90 3.90±5.40 6.70±7.10 10.20±9.90 4.10 [0.52, 7.68] 0.66 [0.06, 1.26]
Stride time DFA Lamoth, 2011 Word naming
with 'R' or 'G' 0.87±0.15 0.74±0.15 0.84±0.16 0.84±0.11 0.13 [0.01, 0.25] 0.81 [0.01, 1.62]
SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; DFA=Detrended fluctuation analysis; *=ES was
reversed to indicate change in the direction of other outcome measures. ESs were calculated as standardized mean difference and 95% CI.
-90-
TABLE 2.7 Gait outcome: within study MCI vs. dementia
Author, Year [reference] Task
Control group Cognitive impairment
group Mean Difference (95% CI)
Between group ES (95% CI) Single-task
Mean±SD Dual-task Mean±SD
Single-task Mean±SD
Dual-task Mean±SD
Stride time CV
Muir, 2012
Backwards counting from 100 by 1s
2.59±1.47 4.06±2.37 2.67±1.08 4.86±2.74 0.72 [0.00, 1.44] 0.54 [-0.02, 1.10]
Animal naming 2.59±1.47 8.02±8.88 2.67±1.08 9.04±8.94 0.94 [0.22, 1.66] 0.71 [0.14, 1.27]
Backwards counting from 100 by 7s
2.59±1.47 10.07±9.29 2.67±1.08 12.49±12.33 2.34 [1.62, 3.06] 1.76 [1.11, 2.41]
Step regularity
Gillian, 2009
Backwards counting from 50 by 1s
287.00±29.00 224.00±47.00 227.00±82.00 139.00±81.00 -25.00 [-72.58, 22.58] 0.48 [-0.49, 1.45]*
Step time variance
Konig, 2017
Backwards counting from 305 by 1s
5.70±4.50 6.80±5.30 6.70±7.10 10.20±9.90 2.40 [-0.98, 5.78] 0.40 [-0.18, 0.98]
-91-
MCI=Mild Cognitive Impairment; SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; *=ES was
reversed to indicate change in the direction of other outcome measures. ESs were calculated as standardized mean difference and 95% CI.
-92-
TABLE 2.8 Stratification of outcomes for meta-analysis
Grouping Stratification Variable Tau2 Q df I2 Z ES Single-task vs. Dual-task
Total
1.32 553.30 84 (P<0.00001) 85% N/A SMD
Cognitive status MCI only 2.21 233.29 29 (P<0.00001) 88% N/A SMD
Dementia only 0.84 269.47 54 (P<0.00001) 80% N/A SMD
Cognitive status and gait outcome
MCI and stride time CV only 5.24 96.03 16 (P<0.00001) 83% N/A MD
Dementia and stride time CV only
6.61 112.18 18 (P<0.00001) 84% N/A MD
Cognitive status, gait outcome and dual-task paradigm
MCI, stride time CV and backwards counting by 1s
7.79 90.76 4 (P<0.00001) 96% N/A MD
MCI, stride time CV, categorical verbal fluency
3.39 74.21 4 (P<0.00001) 95% N/A MD
Dementia, stride time CV, Backwards counting by 1s
169.27 4964.94 6 (P<0.00001) 100% N/A MD
Dementia, stride time CV, categorical verbal fluency
17.43 20.38 2 (P<0.0001) 90% N/A MD
Control vs. Cognitively impaired
Total
3.35 20054.93 40 (P<0.00001) 100% N/A SMD
Cognitive status MCI only 0.21 1174.57 18 (P<0.00001) 98% N/A SMD
Dementia only 58.08 14707.80 21 (P<0.00001) 100% N/A SMD
Cognitive status and gait outcome
MCI and stride time CV only 4.46 245.24 8 (P<0.00001) 97% N/A MD
-93-
Dementia and stride time CV only
42.5 3917.86 12 (P<0.00001) 100% N/A MD
Cognitive status, gait outcome and dual-task paradigm
MCI, stride time CV and backwards counting by 1s
11.21 96.88 2 (P<0.00001) 98% N/A MD
MCI, stride time CV, categorical verbal fluency
1.57 23.3 2 (P<0.00001) 91% N/A MD
Dementia, stride time CV, Backwards counting by 1s
78.01 2845.79 4 (P<0.00001) 100% N/A MD
Dementia, stride time CV, categorical verbal fluency
23.33 75.82 2 (P<0.00001) 97% N/A MD
Within study: MCI vs. Dementia
Total 0 3.74 4 (0.44) 0% 3.06 (0.002)
SMD
MCI=Mild Cognitive Impairment; I2=Measures heterogeneity; Z=Test for overall effect; ES=Effect size; SMD=Standardized mean difference;
MD=Mean difference; CV=Coefficient of variance; P=p value, denoting significance.
-94-
CHAPTER 3: STUDY
The Effects of Dual-Tasking on Gait Dynamics and Cognitive Performance in Adults
with Mild Cognitive Impairment: Contributing Factors and Nonlinear Outcomes
Authors and contributions:
Tess C Hawkins MExPhys, Rebecca Samuel BAppSc, Yorgi Mavros PhD, Nicola Gates PhD,
Guy C Wilson MSc, Nidhi Jain MPH, Jacinda Meiklejohn BS, Henry Brodaty DSc, Wei Wen
PhD, Nalin Singh MBBS, Bernhard T Baune PhD, Chao Suo PhD, Michael K Baker PhD,
Nasim Foroughi PhD, Yi Wang PhD, Perminder S Sachdev PhD, Michael J Valenzuela PhD,
Jeffrey M Hausdorff PhD, and Maria A Fiatarone Singh MD.
Study concept and design: TCH, RS, YM, MAFS and JMH. Acquisition of data: NG, GCW,
NJ, JM, CS, MKB, NF and YW. Analysis and interpretation of data: TCH, RS, YM, MAFS
and JMH. Drafting of the manuscript: TCH, RS, YM, JMH and MAFS. Critical revision of
the manuscript for important intellectual content: TCH, RS, NG, GCW, NJ, JM, WW, MKB,
NF, YW, HB, NS, BTB, CS, PSS, MV, YM, MAFS and JMH. Statistical analysis: TCH, RS
and YM. Obtained funding: HB, WW, NS, BTB, PSS, MV, and MAFS. Administrative,
technical, and material support: NJ, JM, CS. Study supervision: NG, HB, PSS, MV and
MAFS.
-95-
Faculty of Health Sciences
3.1 AUTHOR CONTRIBUTION STATEMENT
Candidate Name: Tess C Hawkins
Degree Title: Master of Applied Science (Research)
Paper Title: The effects of dual-tasking on gait dynamics and cognitive performance in
adults with Mild Cognitive Impairment
As the research supervisor of the above candidate, I confirm that the above candidate has
made the following contributions to the above paper title:
- Conception and design of the research
- Analysis and interpretation of the findings
- Writing the paper and critical appraisal of content
Professor Maria Fiatarone Singh
Discipline of Exercise & Sport Science
Faculty of Health Sciences
The University of Sydney
-96-
3.2 PREAMBLE
The preceding chapter (Chapter 2) detailed the dual-task cost of gait performance in
cognitively impaired older adults. The majority of the studies (92%) reported linear gait
outcome measures only, with nonlinear gait outcome measures reported in just two studies
(8%). The low representation of nonlinear gait dynamics data may distort the understanding
of how impactful dual-tasking is on cognitively impaired older adults. Additionally, the
reporting of characteristics associated dual-task gait dynamics was lacking, which limits the
understanding of what contributes to dual-task cost in cognitively impaired older adults. This
chapter (Chapter 3) presents the findings of a study that used both linear and nonlinear
measures to determine the effects of dual-tasking on gait dynamics in adults with cognitive
impairment, and identified physical, psychosocial, and structural brain characteristics
associated with dual-task cost.
-97-
3.3 ABSTRACT
Objectives
Individuals with Mild Cognitive Impairment (MCI) have increased gait variability. Dual-
tasking can detect interactions between gait dynamics, including variability, and cognition.
We aimed to determine the acute effects of dual-tasking on gait dynamics and cognition in
MCI and to identify associated clinical characteristics.
Design
Acute exposure to dual-tasking during baseline assessment of a randomized controlled trial.
Setting and participants
Ninety-three individuals with MCI (mean age 70±6.8 years; 66.6% female) from the
interventional ‘SMART Study’.
Methods
Cognition, gait, brain Magnetic Resonance Imaging (MRI), muscle strength, aerobic
capacity, body composition, physical and psychosocial function were assessed. Dual-task gait
was measured using force-sensitive insoles to quantify temporal gait dynamics; specifically,
stride time variability and DFA (detrended fluctuation analyses fractal scaling exponent). The
relationship between gait dynamics and cognitive performance was evaluated using linear
mixed models with repeated measures, adjusted for confounders. Linear regression explored
hypothesized mediators of the potential dual-tasking deficits.
Results
Gait dynamics worsened significantly during dual-tasking, with performance decrements in
both stride time variability (p<0.001) and DFA (p=0.001). Lower aerobic capacity and thinner
-98-
posterior cingulate cortex were associated with greater performance decrements in DFA;
whereas smaller hippocampal volume, worse psychological well-being and poorer static
balance were associated with greater performance decrements in stride time variability.
Notably, cognitive performance on the secondary dual-task did not change under dual-task
conditions.
Conclusions/implications
Participants with MCI preserved their cognitive performance at the cost of their gait dynamics
when dual-tasking. We have shown, for the first time that the decrements in dual-tasking gait
are associated with lower aerobic fitness, balance, psychological well-being, and brain
volume in cognitively-relevant areas of the posterior cingulate and hippocampus in MCI; all
of these characteristics are modifiable by exercise. Thus, targeted exercise interventions are
needed to determine the potential plasticity of gait dynamics when stressed in vulnerable
cohorts.
Key words
Dual-task, Gait variability, Gait dynamics, Mild cognitive impairment, and Walking.
-99-
3.4 INTRODUCTION
The prevalence of gait disorders increases with age, affecting up to 35% of community-
dwelling older adults [1]. Increasing gait variability is associated with increased risk of falls
[2] and reduced mobility [3]. Mild Cognitive Impairment (MCI), an intermediate stage
between normal cognition and dementia, is associated with greater gait variability [4] and
double the risk of injurious and multiple falls compared to cognitively normal adults [5].
Dual-tasking is a sensitive method used to investigate interactions between gait variability
and cognitive domains [6, 7], and is associated with increased fall risk amongst older adults
[8]. Dual-tasking impairs gait in individuals with deficits in cognitive function, including
MCI and Alzheimer’s disease [7].
Despite the well-characterized worsening of gait under dual-task conditions, studies
evaluating dual-task associations in older adults with MCI are limited. Identification of
modifiable characteristics associated with gait dynamics under dual-task conditions may lead
to targeted interventions to reduce falls in older adults with MCI. Therefore, we aimed to
determine the effects of dual-tasking on gait dynamics and cognitive performance in adults
with MCI, and to identify physical, psychosocial, and structural brain characteristics
associated with decrements due to dual-tasking (dual-task cost). We hypothesized that there
would be a worsening of both gait and the performance of the secondary cognitive task during
the dual-task condition, and that these reductions would be associated with lower strength,
aerobic capacity, functional performance, psychosocial function and smaller hippocampal
volume and posterior cingulate cortex thickness. These factors were selected a priori due to
their known decrements in cognitive impairment, frailty, or falls [9].
3.5 METHODS
-100-
The complete study protocol for the Study of Mental and Resistance Training (SMART) has
been published [10], with primary [11] and secondary outcomes published [12, 13]. The study
was approved by the Royal Prince Alfred Human Research Ethics Committee (X04-0064),
and written informed consent was obtained from all participants. The study was registered
with the Australia New Zealand Clinical Trials Registry (ACTRN12608000489392).
Participants
One hundred community-dwelling older adults with MCI (Peterson criteria [14]) were
recruited. Two participants could not wear gait monitors due to fused toes, while technical
issues, including incomplete data recording, precluded full gait data in five others. Thus, gait
dynamics data were available for 93 participants. All participants were assessed within the
research clinic space at The University of Sydney Cumberland Campus at Lidcombe in New
South Wales, Australia.
Assessment of cognitive function
Baseline cognitive function has been published [10]. The primary outcome of the SMART
trial was global cognition assessed using the Alzheimer’s Disease Assessment Scale -
Cognition (ADAS-Cog). Attention/speed was assessed via Symbol Digit Modalities Test
(SDMT) and Trail Making Test A. Executive function was assessed by Matrices and
Similarities subtests of the Wechsler Adult Intelligence Scale 3rd Edition (WAIS-III) and
verbal fluency (Controlled Oral Words Association Test (COWAT) and Animal Naming) and
the difference between Trail Making Tests B and A (Trails B – A). Memory tests included
auditory Logical Memory I (immediate) and II (delayed) subtests of the Wechsler Memory
Scale 3rd Edition (WMS-III) and the List Learning subsection of the ADAS-Cog, and visual
via Benton Visual Retention Test-Revised 5th Edition (BVRT-R). Global Domain was the
-101-
average of all z-scores for the gait sample participants (n=93). This included all tests except
List Learning, as it was already included within ADAS-Cog total score.
Assessment of letter fluency at rest and during ambulation
Letter fluency, a subcategory of verbal fluency, was assessed using the COWAT [15] to
measure cognitive performance. Participants were instructed to name as many words
beginning with the letter “F” in 1 minute (FSINGLE), excluding proper nouns, repeated words,
and variations of the same word using a prefix or suffix (e.g., bath and bathing). A score was
calculated by totaling the number of admissible words.
Assessment of gait dynamics
Gait dynamics were assessed similarly to methods previously described [2, 16, 17]. Briefly,
force-sensitive insoles were placed in the participants’ shoes to measure the force applied to
the ground during ambulation. Two sensors were used, one under the heel and another under
the forefoot and toes. A small, lightweight recorder with an on-board A/D converter (12 bit)
was worn on the ankle to sample the output of the insoles at 300Hz and record the data. The
digitized data were transferred to a workstation for analysis, using software that extracts the
initial contact time of each stride [17].
Participants were instructed to walk at their preferred walking speed for two minutes in a
well-lit, indoor hallway with an open path of 25 meters in length. When at the end of the
hallway a large half circle turn was instructed rather than turning on the spot. The first and
last ten seconds of each assessment were removed to minimize starting acceleration or ending
deceleration effects, and a median filter was applied to data points that were three SDs higher
or lower than the median value to remove any outliers due to turns or other irregular gait
-102-
patterns [16]. Participants were instructed to wear habitual, low-heeled shoes, comfortable
clothing, visual and hearing aids, if required.
Gait assessment was performed under two conditions, ‘single-task’ (undistracted walking)
and ‘dual-task’ (distracted walking), in a randomized order with no instruction about task
priority. Stride time variability (coefficient of variation, CV) and detrended fluctuation
analyses fractal scaling exponent (DFA) [18-20] were calculated for both single-task
[(CVSINGLE) and (DFASINGLE)] and dual-task [(CVDUAL) and (DFADUAL)] conditions. To
assess stride-to-stride variability and arrhythmicity of gait, the coefficient of variation (CV)
in each participant’s stride time was calculated using the formula CV = (Standard Deviation
/ Mean) * 100, with lower CV indicating more stable gait. To quantify how the dynamics of
stride times fluctuate and change over time, we applied detrended fluctuation analysis (DFA)
to each participant’s sequence of stride times. Specifically, DFA is a scaling analysis method
used to quantify long-range power-law correlations in signals, and evaluates the fractal
scaling of exponents and the degree of randomness in highly non-stationary physiological
data. DFA eliminates trends in time-series, and can therefore avoid the spurious detection of
correlations from non-stationary artefacts. In general, physiologically healthy systems have
fractal scaling indices between 0.8 and 1.0, with values closer to 0.5 indicating a less healthy
state [21]. The importance of this index is supported by findings among adults with known
altered gait dynamics, in whom a lower fractal scaling index was the only gait dynamics
parameter found to distinguish fallers from non-fallers [21]. The CV and DFA assess and
quantify the changes over time in gait through different methods, the former captures the
magnitude of the changes and the latter captures changes over time. Both are necessary to
assess the totality of gait dynamics and associated factors.
-103-
The secondary task was the COWAT test, with the number of correct “F words” counted in
the first minute of walking (FDUAL). The letter “F” was selected to remain consistent with the
seated COWAT test described above (FSINGLE). No instruction was provided about
prioritizing performance of either the walking or cognitive task but they were encouraged to
do their best during all trials. The words were recorded using a portable recorder (Samsung
YP-U3, Samsung Electronics Co., South Korea) and the recording was subsequently
reviewed by the research assistant who counted the number of admissible words during the
first and second minute. The number of correct words in the first minute of the dual-task gait
condition (FDUAL) was compared to the number of correct words during the FSINGLE condition.
The order was not randomized for the assessment of “F” words, with the dual-task condition
performed 1 week after the seated COWAT. The ‘dual-task cost’ for CV, DFA and F words
were calculated (CVCOST= CVDUAL – CVSINGLE; DFACOST= DFADUAL – DFASINGLE; FCOST =
FDUAL – FSEINGLE).
Assessment of Neuroimaging Outcomes
Details of the neuroimaging assessment outcomes have been published [10, 13]. Magnetic
Resonance Imaging (MRI) data were acquired using a 3.0-Tesla Philips Achieva System
(Achieva, Phillips Healthcare, Best, The Netherlands). Brain structure was assessed using a
T1-weighted whole brain scan (sequence: T1TFE; TR/TE: 6.39/2.9 ms; slice thickness 1.0
mm without gap; field of view: 256 × 256; resolution 1 × 1 mm). Brain volumes were assessed
using 1H-MRS for regional measures: left hippocampus (20 mm M/L, 15 mm D/V, 30 mm
A/P, oriented along the hippocampus) and posterior cingulate grey matter (20 mm M/L, 20
mm D/V, 20 mm A/P) using the PRESS sequence (TE/ TR = 30/2000 ms, 1024 points, 256
averages). Automated and semi-automated computational neuroanatomical analyses were
-104-
performed using a combination of different software packages, in addition to expert manual
tracing of hippocampus.
Assessment of peak strength
Maximal strength testing was assessed via the one-repetition maximum (1RM) on Keiser
pneumatic resistance machines (Keiser Sports Health Equipment, Ltd., Fresno, CA).
Participants’ 1RM was determined on the leg press, knee extension, hip abduction, chest press
and seated row machines. One RM tests were performed twice, one week apart, with the best
performance used.
Assessment of peak aerobic capacity
Aerobic capacity (VO2peak) was determined via indirect calorimetry during a physician-
administered, graded treadmill walking test to volitional fatigue. Methods and data handling
have been previously published [12].
Measures of physical function
Details of the physical function assessments have been published [10]. Static balance was
assessed on one attempt using six different positions (wide stance, narrow stance, semi-
tandem stance, tandem stance, one leg with eyes open and one leg with eyes closed). The
time achieved for each stance was measured, with participants instructed to maintain balance
for 15 seconds. The total static balance score (maximum 90) was calculated by summing the
time for all six positions [22]. All other physical function testing was done in duplicate, with
the better score used in analyses except for habitual gait velocity, for which the average was
used. Habitual and maximal gait velocities were assessed over two metres using an Ultra-
timer (Raymar, Oxfordshire, UK). Participants were instructed to start walking, with the timer
-105-
initiated after the participant had walked 2-meters and the timer automatically stopping as
soon as the participant had walked a further 2-meters. The participant was instructed to walk
to an endpoint 3-meters beyond where the timing would stop. This process was implemented
to avoid the recording of acceleration and deceleration. Dynamic balance was assessed using
the time taken to forward tandem walk over a 3-meter marked course. The instruction to the
participant was to walk as quickly as possible and with as few errors as possible. Errors
included touching the examiner or any object in the assessment environment, stepping
without heel-toe contact and losing balance requiring the support of the examiner. Time was
only stopped when the participant’s whole foot crossed the 3-meter line. Lower extremity
function and power was assessed using the sit-to-stand [23] and stair climb tests [10]. For the
sit-to-stand, participants were instructed to stand up and sit down 5 times as quickly as
possible. Finally, stair climb power (W) was measured by asking participants to ascend a
flight of stairs as quickly as possible. Stair climb power was calculated using the following
formula P (watts) = (M × D) × 9.8/t Where: M = Body mass (kg), D= Vertical distance (m),
D = vertical height of the staircase, and t = Time (s). The 6-minute walk test (6MWT) was
assessed twice, at least one week apart.
Anthropometry and Body Composition
Height, naked body mass and waist circumference were measured as the mean of triplicate
measures after a 12-hour overnight fast. Height was measured to the nearest 0.1cm using
stretch stature with a wall mounted Holtain stadiometer (Holtain Limited, Crymmych Pembs.,
UK). Body mass was measured using a calibrated scale HW-100k & SECA Wedderburn
(>100 kg). Body Mass index (BMI) was calculated by dividing the participants’ body mass
in kilograms by the square of their height in metres. Waist circumference was measured with
Lufkin steel tape measure (W606 PM), using the International Diabetes Federation (IDF)
-106-
protocol [24]. Bioelectrical Impedance Analysis (BIA; RJL Systems, Inc., Clinton, MI, USA)
was used to evaluate body composition. Whole body skeletal muscle mass (kg) [25] and fat
free mass (kg) [26] were calculated using the average resistance and reactance values of three
sequential BIA measures. Further details of the assessments are published [10].
Measures of Psychosocial Function
Details of the assessments used to measure psychosocial function and assess psycho-social
well-being and quality of life have been published [10]. The following tests were used to
measure psychosocial function and assessed psycho-social wellbeing and quality of life via
the Life Satisfaction Scale (LSS) [27], Scale of Psychological Well Being (SPWB) [28],
Quality of Life Scale (QOLS) [29], Physical and Mental Health Short-36 (SF-36), Depression
Anxiety Stress Scale (DASS 21) [30], Memory Awareness Rating Scale – Memory
Functioning Scale (MARS-MF) [31], Duke Social Support Index Scale (DSSIS) [32] and
Life Experience Questionnaire (LEQ).
Statistical Methods
Data were inspected for normality. Normally-distributed data are presented as mean±SD and
non-normally distributed data presented as median (interquartile range). All CV (CVSINGLE,
CVDUAL and CVCOST) variables were log-transformed prior to use in parametric statistics.
Sequential linear regression models (adjusted for age and sex) were constructed including
potential confounders associated with the number of F words during the COWAT test
(FSINGLE) and single- and dual-task gait dynamics (CVSINGLE and DFASINGLE). Next, linear
mixed models with repeated measures were constructed to determine the effect of dual-
tasking on both letter fluency and gait dynamics. The single-task condition (FSINGLE, CVSINGLE
and DFASINGLE) was entered as time point 1, and the ‘dual-task’ condition (FDUAL, CVDUAL
-107-
and DFADUAL) as time point 2. A compound symmetry covariance matrix was used. Models
were adjusted for age and sex, with gait dynamics outcomes further adjusted for the order of
the walking condition. Next, linear regression models were constructed to determine variables
associated with FCOST, CVCOST and DFACOST. Models were adjusted for age, sex, and baseline
score of the dependent variable and order of the condition for the gait dynamics assessment.
All models involving cognitive performance (including FSINGLE and FCOST) were further
adjusted for education, while models involving DFA were further adjusted for the number of
medications, a covariate. Statistical significance was assumed at <0.05 level without
Bonferroni adjustment, as all hypotheses were specified a priori. [33]. All data were analysed
using IBM SPSS (version 24; IBM Corp., Armonk, NY).
3.6 RESULTS
Baseline clinical characteristics of the participants have been published [11]. Data for the
available 93 participants (66.6% women), were similar to the overall cohort (p>0.05 for all
variables). The average age was 70.0±6.8 years, MMSE score 27.5±1.4, and habitual gait
speed 1.21±0.24 m/s, with 17% of participants having a gait speed below 1.0 m/s. Participant
characteristics data are presented in Table 3.1.
Factors associated with gait dynamics
The number of medications prescribed was inversely associated with DFASINGLE (r=-0.23,
p=0.029), but not CVSINGLE (r=-0.03, p=0.750). There were no associations between age, sex,
years of education, smoking status, drinking status or the number of chronic diseases and
either index of gait dynamics. Consequently, all analyses with DFACOST as a dependent
variable were adjusted for the number of medications.
-108-
Dual-Task Cost
Data are presented in Figure 3.1. As hypothesized, gait dynamics worsened significantly
during dual-tasking, with decrements in performance observed for stride time variability
(single-task 2.012 (0.767), dual-task 2.555 (2.227)) and DFA (single-task 0.804±0.151, dual-
task 0.745±0.160). However, contrary to our hypothesis, cognitive performance on the
COWAT did not significantly change under dual-task conditions (single-task 13±5, dual-task
12±4 (Figure 3.1). Changes in letter fluency (FCOST) were not associated with changes in gait
dynamics for either CVCOST (r=-0.14, p=0.171) or DFACOST (r=0.03, p=0.746).
Factors associated with changes in stride time variability during dual-tasking (CVCOST)
Data are presented in Table 3.2. Contrary to our hypotheses, cognitive performance
(executive function, memory, attention and global domains) was not associated with CVCOST
under dual-task conditions (p>0.05). However, higher CVCOST was associated with smaller
(or lower) left hippocampus volume (r=-0.35, p=0.023) with a similar trend for total
hippocampus volume (r=-0.31, p=0.050). Gait dynamics and brain morphology data are
presented in Figure 3.2.
Worse dynamic balance (longer tandem walk time) was directly associated with a higher
CVCOST (r=0.28, p=0.022). However, CVCOST was not associated with body composition,
aerobic capacity, strength or other measures of functional performance.
As hypothesized, psychological well-being was inversely associated with CVCOST, with
higher Environmental Mastery, Personal Growth, Personal Relations, Purpose in Life, Self-
acceptance and total score on the Psychological Wellbeing Scale associated with preservation
of gait under dual-task conditions (p<0.05). Similarly, CVCOST was inversely associated with
-109-
Duke Social Support Index Scale (DSSIS), indicating that the lower level of, and satisfaction
with, social support was associated with a greater gait impairment during dual-task conditions
(p<0.05).
Factors associated with changes in fractal scaling exponent of gait during dual-tasking
(DFACOST)
Contrary to our hypotheses, cognition and psychosocial function were not associated with
DFACOST under dual-task conditions (p>0.05). Notably, as anticipated, greater left (r=0.23,
p=0.026) and total (r=0.25, p=0.015) posterior cingulate cortex thickness as well as better
performances in the 6MWT (r=0.25, p=0.025) and aerobic capacity (r=0.24, p=0.033) were
related to preservation of gait during dual-tasking. However, DFACOST was unrelated to lower
body strength (r=0.25, p=0.063), whole body strength (r=0.25, p=0.079), static balance time
(r=0.18, p=0.075) or tandem walk score (r=-0.12, p=0.254).
3.7 DISCUSSION
The primary finding from this investigation was that in older adults with MCI, resting
cognitive performance on a letter fluency task was preserved under dual-task conditions,
whereas a significant worsening of gait dynamics, both in the magnitude (CV) and time
course (DFA), was observed. Furthermore, worsening of DFA was found to be associated
with lower posterior cingulate cortex thickness and aerobic and walking capacity, while
worsening of stride time variability was associated with smaller hippocampal volume, static
balance and psychological well-being. In contrast, single-task (seated) cognitive function
predicted neither gait variability nor gait dynamic changes during dual-tasking. Although
previous studies [4, 34] have focused on the decrements in gait under dual-task conditions,
-110-
our study has investigated the effects of dual tasking on cognitive performance in MCI, as
well as the mediating role of brain morphology and physical/psychological function in MCI.
We observed preservation of cognitive performance while sacrificing gait dynamics under
dual-task conditions in adults with MCI, which is in agreement with previous studies in
healthy older adults [35] and adults with Parkinson’s disease [36]. This may in part explain
why worsening of gait has previously been associated with recurrent fallers [37] and future
risk of falls [2, 38], and why people with MCI are at a greater risk of falling than their healthy
counterparts [5].
Potential mediators of dual-tasking deficits
Brain morphology
We have reported for the first time that thicker left and total posterior cingulate cortices were
associated with preservation of DFA during dual-tasking. The posterior cingulate cortex
contributes to bilateral lower limb coordination [39] and motor imagery [40], thus, posterior
cingulate thickness reduction with age or MCI could theoretically reduce walking
coordination and result in a more variable gait, particularly during dual-tasking [41], as we
observed in relation to DFA. Additionally, the posterior cingulate cortex assists in directing
the focus of attention [42], which requires information to be integrated [43]. Reduction in
thickness may impact the ability of the posterior cingulate to function; indeed individuals
with clinical disorders associated with posterior cingulate cortex abnormalities have difficulty
regulating the focus of attention [43]. Notably, we have shown that high intensity strength
training increases posterior cingulate thickness [13] and that this change in posterior cingulate
thickness is directly associated with the cognitive benefits of the strength training [13]. Future
investigations are needed to determine whether gait dynamics also improve after robust
-111-
strength training, and whether such changes are related to changes in brain morphology or
other adaptations such as muscle strength or balance.
We also report for the first time the novel finding that lower left and total hippocampal
volume are associated with greater deficits in stride time variability during dual-tasking.
Individuals with MCI have both a higher stride time variability [44, 45] and a lower
hippocampal volume [44] than cognitively healthy individuals. Interestingly, higher stride
time variability has been significantly associated with a lower hippocampal volume in
cognitively healthy individuals during uninterrupted walking [44], however, unexpectedly
not in individuals with MCI [44]. The hippocampus contributes to rhythmicity of locomotion
[46] and is known to atrophy in individuals with MCI faster than in healthy adults [47], which
supports our observed association between lower hippocampal volume and worse gait
dynamics during dual-tasking. It has been postulated that greater hippocampal volume may
compensate for impaired gait dynamics [39] and the hippocampal atrophy of MCI may
mediate diminished ability to regulate gait dynamics under the stressful condition of dual-
tasking in this cohort.
Thus, our novel results for both posterior cingulate cortex and hippocampus provide strong
new evidence of associations between brain morphology, cognitive impairment and gait
variability [48]. We have previously shown that resistance training can significantly increase
posterior cingulate cortex thickness in MCI, and that this increase is associated with the
cognitive benefits of the exercise [13]. This suggests potent avenues to investigate
mechanisms by which exercise may improve gait dynamics in vulnerable cohorts with
cognitive or other neurological impairment.
-112-
Physical fitness
We have shown that higher aerobic capacity and 6-min walking distance are associated with
better preservation of DFA during dual-tasking. High intensity treadmill training has been
shown to improve aerobic capacity and cognition in adults with amnestic MCI [49], whereas
most other low-moderate intensity aerobic interventions have yielded non-significant
findings in this cohort [50]. Progressive resistance training can also improve aerobic capacity
and cognitive function, (mediated by muscle strength gains), as we have previously shown
within this cohort [12]. Notably, given the cross-sectional nature of our analyses, reverse
causality cannot be excluded. Individuals with a less variable gait may be more likely to walk
more frequently and thus have a greater aerobic capacity and functional performance
compared to individuals with a more variable gait pattern. Longitudinal exercise studies are
required to investigate whether improved aerobic/walking capacity will also improve gait
variability during single- and dual-task conditions.
Muscle strength was not related to the preservation of either DFA or stride time variability
during dual-task walking. This differs from data we previously reported showing that higher
muscle strength was related to lower variability measures in community-dwelling older adults
[2] and increased strength predicted less variability in older adults with mild functional
impairment after a multi-modal exercise intervention [51]. Alternatively, others have reported
that stride time variability in older adults with higher level gait disorders was not associated
with muscle strength [21], nor was step time variability improved by a muscle strengthening
intervention [52]. The above studies only measured gait variability under single-task
conditions, hence making this study the first to investigate the relationship between strength
and gait variability under dual-task conditions. Herman and colleagues [21] speculated that
the lack of association between strength and gait variability in some prior studies was due to
-113-
the origin of gait variability being pathologic rather than motor-based, suggesting that frontal
lobe dysfunction not modifiable by strength training may be a cause of gait variability rather
than muscle weakness, which may support our findings.
Psychological well-being
The dual-task cost of stride time variability was greater in those with lower psychological
well-being (SPWB) across the domains of Environmental Mastery, Personal Growth,
Personal Relations, Purpose in Life and Self-acceptance. These data are in agreement with
previous evidence suggesting that worse stride time variability is associated with an increased
fear of falling [21] and fear of falling has been previously associated with physical and mental
limitations as well as social functioning [53]. Consistent with our findings on psychological
well-being, depression is a well-known risk factor for falls and hip fractures [54], independent
from anti-depressant medications, which pose additional fall risk [55]. Whether gait dynamics
and brain morphology decrements in depression underlie this risk is an important area for
future investigation, given the known relationship between hippocampal atrophy and
depression [56] and that our cohort were free from major depression. Additionally, MCI has
been associated with reduced psychological wellbeing [57] and increased falls [38],
supporting a potential link between worsened gait dynamics and poor psychological
wellbeing, but again reverse causality or bi-directional relationships cannot be ruled out.
Limitations
As noted above, reverse causality could explain some of the study outcomes due to the cross-
sectional study design. Also, as noted, the order of the COWAT assessment and dual-task
walking was not randomized. The presentation of the “F” word task at rest prior to dual-
tasking may have produced a learning effect, minimizing the observation of cognitive deficits
-114-
during dual-tasking. Future studies should randomize the sequencing of single- and dual-task
cognitive performance. Finally, appropriately designed studies are warranted to determine
which clinical characteristics related to the dual-task cost remain as independent predictors
following multiple or stepwise regression.
3.8 CONCLUSIONS AND IMPLICATIONS
Older adults with MCI preserved their cognitive performance at the cost of the variability and
dynamics of their gait under dual-task conditions. Novel associations were observed between
worsening of the fractal scaling exponent of gait and posterior cingulate cortical thickness,
while worsening of stride time variability was associated with lower hippocampal volume.
Better aerobic and walking capacity, psychological wellbeing and static balance were also
associated with preservation of gait during a cognitive stressor. Notably, all these factors have
previously been shown to be modifiable with robust exercise modalities in clinical trials.
Thus, longitudinal research is required to determine the extent to which gait dynamics are
also modifiable, and the optimal exercise prescriptions needed to promote optimization of
gait patterns, and ultimately reduce fall risk in vulnerable cohorts.
-115-
3.9 REFERENCES
1. Verghese, J., et al., Epidemiology of Gait Disorders in Community-Residing Older
Adults. Journal of the American Geriatrics Society, 2006. 54(2): p. 255-261.
2. Hausdorff, J.M., D.A. Rios, and H.K. Edelberg, Gait variability and fall risk in
community-living older adults: A 1-year prospective study. Archives of Physical
Medicine and Rehabilitation, 2001. 82(8): p. 1050-1056.
3. Brach, J.S., et al., Gait variability and the risk of incident mobility disability in
community-dwelling older adults. The Journals of Gerontology Series A: Biological
Sciences and Medical Sciences, 2007. 62(9): p. 983-988.
4. Montero-Odasso, M., S.W. Muir, and M. Speechley, Dual-task complexity affects gait
in people with mild cognitive impairment: the interplay between gait variability, dual
tasking, and risk of falls. Archives of Physical Medicine & Rehabilitation, 2012.
93(2): p. 293-299.
5. Delbaere, K., et al., Mild Cognitive Impairment as a Predictor of Falls in Community-
Dwelling Older People. The American Journal of Geriatric Psychiatry, 2012. 20(10):
p. 845-853.
6. Bahureksa, L., et al., The Impact of Mild Cognitive Impairment on Gait and Balance:
A Systematic Review and Meta-Analysis of Studies Using Instrumented Assessment.
Gerontology, 2017. 63(1): p. 67-83.
7. Bragatto, V.S.R., et al., Dual-task during gait between elderly with mild cognitive
impairment and Alzheimer: systematic review. Fisioterapia em Movimento, 2017. 30:
p. 849-857.
8. Beauchet, O., et al., Stops walking when talking: A predictor of falls in older adults?
European Journal of Neurology, 2009. 16(7): p. 786-795.
9. Clegg, A., et al., Frailty in elderly people. The lancet, 2013. 381(9868): p. 752-762.
10. Gates, N.J., et al., Study of Mental Activity and Regular Training (SMART) in at risk
individuals: A randomised double blind, sham controlled, longitudinal trial. BMC
Geriatrics, 2011. 11(1): p. 19.
-116-
11. Fiatarone Singh, M.A., et al., The Study of Mental and Resistance Training (SMART)
study—resistance training and/or cognitive training in mild cognitive impairment: a
randomized, double-blind, double-sham controlled trial. Journal of the American
Medical Directors Association, 2014. 15(12): p. 873-880.
12. Mavros, Y., et al., Mediation of Cognitive Function Improvements by Strength Gains
After Resistance Training in Older Adults with Mild Cognitive Impairment: Outcomes
of the Study of Mental and Resistance Training. Journal of the American Geriatrics
Society, 2017. 65(3): p. 550-559.
13. Suo, C., et al., Therapeutically relevant structural and functional mechanisms
triggered by physical and cognitive exercise. Molecular Psychiatry, 2016: p. doi:
10.1038/mp.2016.19.
14. Petersen, R.C., Mild Cognitive Impairment: Clinical Characterization and Outcome.
The Journal of the American Medical Association, 1999. 281(19): p. 303-308.
15. Kaufman, A.S. and E.O. Lichtenberger, Essentials: of WAIS-III assessment. 1999,
New York: J. Wiley & Sons.
16. Hausdorff, J.M., et al., Increased gait unsteadiness in community-dwelling elderly
fallers. Archives of Physical Medicine and Rehabilitation, 1997. 78(3): p. 278-283.
17. Hausdorff, J.M., Z. Ladin, and J.Y. Wei, Footswitch system for measurement of the
temporal parameters of gait. Journal of Biomechanics, 1995. 28(3): p. 347-351.
18. Hausdorff, J.M., Gait dynamics, fractals and falls: Finding meaning in the stride-to-
stride fluctuations of human walking. Human Movement Science, 2007. 26(4): p. 555-
589.
19. Hausdorff, J.M., et al., Is walking a random walk? Evidence for long-range
correlations in stride interval of human gait. Journal of Applied Physiology, 1995.
78(1): p. 349-358.
20. Hausdorff, J.M., et al., Fractal dynamics of human gait: stability of long-range
correlations in stride interval fluctuations. Journal of Applied Physiology, 1996.
80(5): p. 1448-1457.
-117-
21. Herman, T., et al., Gait instability and fractal dynamics of older adults with a
“cautious” gait: why do certain older adults walk fearfully? Gait & Posture, 2005.
21(2): p. 178-185.
22. Rossiter-Fornoff, J.E., et al., A cross-sectional validation study of the FICSIT common
data base static balance measures. Frailty and Injuries: Cooperative Studies of
Intervention Techniques. The journals of gerontology. Series A, Biological sciences
and medical sciences, 1995. 50(6): p. M291.
23. Guralnik, J.M., et al., A short physical performance battery assessing lower extremity
function: Association with self-reported disability and prediction of mortality and
nursing home admission. Journals of gerontology, 1994. 49(2): p. M85-M94.
24. Alberti, K.G., P.Z. Zimmet, and J.M. Shaw, Metabolic syndrome-- a new world-wide
definition. A Consensus Statement from the International Diabetes Federation.
Diabetic medicine : A journal of the British Diabetic Association, 2006. 23(5): p. 469-
480.
25. Jansen, I., et al., Estimation of skeletal muscle mass by bioelectrical impedance
analysis. Journal of Applied Physiology, 2000. 89(2): p. 465-471.
26. Lukaski, H., et al., Validation of tetrapolar bioelectrical impedance method to assess
human body composition. Journal of Applied Physiology, 1986. 60(4): p. 1327-1332.
27. Andrews, F.M. and S.B. Withey, Social indicators of well-being: Americans'
perceptions of life quality. 1976, New York: Plenum Press.
28. Ryff, C.D. and C.L.M. Keyes, The Structure of Psychological Well-Being Revisited.
Journal of personality and social psychology, 1995. 69(4): p. 719-727.
29. Burckhardt, C.S. and K.L. Anderson, The Quality of Life Scale (QOLS): Reliability,
validity, and utilization. Health and Quality of Life Outcomes, 2003. 1(1): p. 60-60.
30. Ng, F., et al., The validity of the 21-item version of the Depression Anxiety Stress
Scales as a routine clinical outcome measure. Acta Neuropsychiatrica, 2007. 19(5):
p. 304-310.
-118-
31. Clare, L., et al., Assessing awareness in early-stage Alzheimer's disease: Development
and piloting of the Memory Awareness Rating Scale. Neuropsychological
Rehabilitation, 2002. 12(4): p. 341-362.
32. Koenig, H.G., et al., Abbreviating the Duke Social Support Index for use in
chronically ill elderly individuals. Psychosomatics, 1993. 34(1): p. 61-69.
33. Nakagawa, S., A farewell to Bonferroni: the problems of low statistical power and
publication bias. Behavioral Ecology, 2004. 15(6): p. 1044-1045.
34. Nascimbeni, A., et al., Dual task-related gait changes in patients with mild cognitive
impairment. Functional Neurology, 2015. 30(1): p. 59-65.
35. Corp, D.T., et al., Reduced motor cortex inhibition and a ‘cognitive-first’
prioritisation strategy for older adults during dual-tasking. Experimental
Gerontology, 2018. 113: p. 95-105.
36. Bloem, B.R., et al., The "posture second" strategy: A review of wrong priorities in
Parkinson's disease. Journal of the Neurological Sciences, 2006. 248(1-2): p. 196-
204.
37. Brach, J.S., et al., Too much or too little step width variability is associated with a fall
history in older persons who walk at or near normal gait speed. Journal of
NeuroEngineering and Rehabilitation, 2005. 2: p. 21-21.
38. Callisaya, M.L., et al., Gait, gait variability and the risk of multiple incident falls in
older people: a population-based study. Age and Ageing, 2011. 40(4): p. 481-487.
39. Tian, Q., et al., The brain map of gait variability in aging, cognitive impairment and
dementia—a systematic review. Neuroscience & Biobehavioral Reviews, 2017. 74: p.
149-162.
40. Rosano, C., et al., Special ArticleGait Measures Indicate Underlying Focal Gray
Matter Atrophy in the Brain of Older Adults. The Journals of Gerontology: Series A,
2008. 63(12): p. 1380-1388.
41. Doi, T., et al., Gray matter volume and dual-task gait performance in mild cognitive
impairment. Brain Imaging and Behavior, 2017. 11(3): p. 887-898.
-119-
42. Leech, R., et al., Fractionating the Default Mode Network: Distinct Contributions of
the Ventral and Dorsal Posterior Cingulate Cortex to Cognitive Control. The Journal
of Neuroscience, 2011. 31(9): p. 3217-3224.
43. Leech, R. and D.J. Sharp, The role of the posterior cingulate cortex in cognition and
disease. Brain, 2014. 137(1): p. 12-32.
44. Beauchet, O., et al., Hippocampal volume, early cognitive decline and gait variability:
Which association? Experimental Gerontology, 2015. 61: p. 98-104.
45. Muir, S.W., et al., Gait assessment in mild cognitive impairment and Alzheimer's
disease: The effect of dual-task challenges across the cognitive spectrum. Gait and
Posture., 2012.
46. Zimmerman, M.E., et al., MRI- and MRS-Derived Hippocampal Correlates of
Quantitative Locomotor Function in Older Adults. Brain Research, 2009. 1291: p. 73-
81.
47. Leal, S.L. and M.A. Yassa, Perturbations of neural circuitry in aging, mild cognitive
impairment, and Alzheimer's disease. Ageing Research Reviews, 2013. 12(3): p. 823-
831.
48. Hamacher, D., et al., Brain activity during walking: A systematic review.
Neuroscience and Biobehavioral Reviews, 2015. 57: p. 310-327.
49. Baker, L.D., et al., Effects of Aerobic Exercise on Mild Cognitive Impairment: A
Controlled Trial. Archives of Neurology, 2010. 67(1): p. 71-79.
50. Gates, N., et al., The effect of exercise training on cognitive function in older adults
with mild cognitive impairment: a meta-analysis of randomized controlled trials. The
American Journal of Geriatric Psychiatry, 2013. 21(11): p. 1086-1097.
51. Hausdorff, J.M., et al., Etiology and modification of gait instability in older adults: a
randomized controlled trial of exercise. Journal of Applied Physiology, 2001. 90(6):
p. 2117-2129.
52. Schwenk, M., et al., Improvements in gait characteristics after intensive resistance
and functional training in people with dementia: a randomised controlled trial. BMC
Geriatrics, 2014. 14(1): p. 73.
-120-
53. Meulen, E., et al., Effect of Fall‐Related Concerns on Physical, Mental, and Social
Function in Community‐Dwelling Older Adults: A Prospective Cohort Study. Journal
of the American Geriatrics Society, 2014. 62(12): p. 2333-2338.
54. Whooley, M.A., et al., Depression, Falls, and Risk of Fracture in Older Women.
Archives of Internal Medicine, 1999. 159(5): p. 484-490.
55. Kvelde, T., et al., Depressive symptoms increase fall risk in older people, independent
of antidepressant use, and reduced executive and physical functioning. Archives of
Gerontology and Geriatrics, 2015. 60(1): p. 190-195.
56. Sheline, Y.I., et al., Hippocampal atrophy in recurrent major depression. Proceedings
of the National Academy of Sciences, 1996. 93(9): p. 3908-3913.
57. Gates, N., et al., Psychological well-being in individuals with mild cognitive
impairment. Clinical Interventions In Aging, 2014. 9: p. 779-792.
-121-
FIGURE LEGENDS FIGURE 3.1: Graphed results under single-task and dual-task walking conditions for (a)
DFA, (b) Stride time variability and (c) Letter fluency. Data for (a) and (c) are presented as
mean±SD and data for (b) are presented as median (IQR) with all adjusted pairwise. DFA is
detrended fluctuation analysis fractal scaling exponent.
FIGURE 3.2: Graphed results of gait dynamics and MRI associations for (a) CVCOST and left
hippocampal volume, (b) DFACOST and left posterior cingulate thickness, and (c) DFACOST
and total posterior cingulate thickness. CVCOST is the dual-task cost of stride time variability.
DFACOST is the dual-task cost of DFA, detrended fluctuation analysis fractal scaling
exponent. Cost variables are calculated by subtracting the single-task from the dual-task.
-122-
FIGURE 3.1 Gait dynamics and cognitive performance under single-task and dual-task
walking conditions
-123-
Data for (a) and (c) are presented as mean±SD and data for (b) are presented as median (IQR)
with all adjusted pairwise. DFA is detrended fluctuation analysis fractal scaling exponent.
Higher DFA and lower CV represent a less variable gait.
-124-
FIGURE 3.2 Relationship between gait dynamics and brain morphology
-125-
CVCOST is the dual-task cost of stride time variability. DFACOST is the dual-task cost of DFA,
detrended fluctuation analysis fractal scaling exponent. Cost variables are calculated by
subtracting the single-task from the dual-task Greater decrements during dual-tasking are
indicated by more positive CV cost and more negative DFA cost.
-126-
TABLE 3.1 Participant characteristics
Characteristics Value
Demographics
Age (yrs) 70.0 ± 6.8
Sex: Female (%) 66.7
BMI (kg/ m2) 26.87 ± 4.89
Education (yrs) 13 ± 3
Total alcoholic drinks / week 3 (9)
Current smoker (%) 2.2
Ex-smoker (%) 38.7
Health status
Medications/day 4 (4)
Number of chronic diseases (%) 2.9 ± 1.7
Osteoarthritis (%) 71.0
Hypertension (%) 40.9
Diabetes (%) 11.8
Gout (%) 4.3
Depressive episodes in the past 5 years 0 (0)
Habitual gait speed (m/s) 1.22 ± 0.24
Results reported in mean±SD or Median (IQR); SD=Standard deviation; IQR=Interquartile
Range; yrs=years; %=percent; BMI=body mass index; kg=kilogram; m= meter; s=second.
-127-
TABLE 3.2 Factors significantly associated with changes in at least one measure of gait
variability and dynamics during dual-tasking
Associated factors
Demographics
CVCOST DFACOST
Age (yrs) -0.032 -0.201*
Psychosocial Assessments
SPWB Environmental Mastery -0.276* -0.091
SPWB Personal Growth -0.217* -0.156
SPWB Personal Relations -0.258* -0.137
SPWB Purpose in Life -0.248* -0.064
SPWB Self-Acceptance -0.244* -0.002
SPWB Total Score -0.303* -0.114
DSSIS (11-33) -0.243* -0.104
Functional Status and Physical Performance
VO2 peak (mL/kg/min) 0.037 0.242*
Best Tandem Walk (s) 0.275* -0.119
6 Minute Walk Distance (m) 0.050 0.245*
Brain MRI Thicknesses and Volumes
Left Hippocampus Volume (mm3) -0.347* -0.184
Left Posterior Cingulate Cortex Thickness
(mm) 0.240 0.228*
Total Posterior Cingulate Cortex Thickness
(mm) 0.246 0.254*
Results reported as r value; yrs=years; kg=kilograms; m= meters; mL=milliliters;
min=minutes; s=seconds; mm=millimeters mm3=millimeter cubed, SPWB=Scale of
-128-
Psychological Well Being; DSSIS=Duke Social Support Index Scale; *=significant p value
(p<0.05).
-129-
CHAPTER 4: CONCLUSIONS AND IMPLICATIONS
4.1 CONCLUSIONS
This thesis has demonstrated a link between cognitive impairment status and the dual-task
cost of gait dynamics, with a particular focus on modifiable and non-modifiable
characteristics within this population. The systematic review (Chapter 2) indicated that gait
dynamics in cognitively impaired older adults are worse when walking under dual-task
conditions compared to single-task conditions. Additionally, the dual-task cost of gait
dynamics is larger for cognitively impaired older adults than healthy older adults. The acute
dual-task study focused on older adults with MCI (Chapter 3) and showed that cognitive
performance was preserved at the cost of worsening their dual-task gait dynamics.
Furthermore, significant and biologically plausible associations were identified between
dual-task gait performance and aerobic capacity, functional performance, psychosocial
function, and brain morphology.
The systematic review (Chapter 2) was aimed at investigating the effect of dual-task walking
on changes in gait dynamics, with respect to differences between cognitive pathology
diagnosis, dual-task paradigms and gait dynamic outcomes in cognitively impaired older
adults. Additionally, this review aimed to identify potentially modifiable and non-modifiable
characteristics of gait dynamics in cognitively impaired older adults. The review found that
gait dynamics are worsened under dual-task conditions compared to single-task conditions in
older adults with both MCI and dementia. The dual-task costs of gait dynamics are further
increased in cognitively impaired older adults compared to healthy older adults, with greater
decrements in gait dynamics observed with more severe cognitive impairment [i.e.,
dementia/Alzheimer’s disease (AD) vs. Mild Cognitive Impairment (MCI)].
-130-
The review also highlighted major limitations within the current literature. Specifically, the
studies were too varied in their methodology to determine if there is a superior dual-task
protocol for measuring gait or falls risk. Studies commonly reported linear measures of gait
dynamics, however, nonlinear measures were only reported twice. Linear and nonlinear
measures of gait dynamics use different methods to assess gait and different aspects of
changes in the gait cycle over time. Both types of outcomes are necessary to fully understand
gait dynamics and associated factors. Additionally, there are MCI and dementia sub-types,
e.g., amnestic MCI (a-MCI) and non-amnestic MCI (na-MCI), and Alzheimer’s disease with
impaired executive function (IEF), extrapyramidal signs (EPS) or frontotemporal dementia
(FTD). There is limited research within these sub-types, and future research should
investigate differences between sub-types with larger sample sizes and varying levels of task
complexity. Furthermore, falls history and fear of falling were not well documented and thus
definitive conclusions about relationships between falls risk and dual-task gait dynamics were
unable to be drawn.
The acute dual-task study (Chapter 3) was aimed at determining the extent of changes in gait
and the performance of a secondary task during dual-task walking using a linear and nonlinear
measure of gait. As identified by the systematic review, there are only two previous studies
published to our knowledge that used nonlinear measures of gait dynamics to quantify the
cost of a dual-task. Additionally, our study sought to explore associations with dual-task
performance and to identify any characteristics that may be modifiable via preventative
therapeutic strategies. These characteristics were identified prior to the study, based on the
existing literature in healthy older adults, gait performance and dual-task walking [1-5], and
included strength, aerobic capacity, functional performance, psychosocial function, and brain
-131-
morphology. Identifying such characteristics associated with dual-task gait performance may
assist in the development of optimal interventions to reduce gait dynamics under single- and
dual-task conditions. The results showed that larger hippocampal volume, thicker posterior
cingulate cortex, higher aerobic and walking capacity, greater psychological wellbeing and
greater dynamic balance were each associated with preservation of gait during dual-task
conditions. These factors have previously been shown to be modifiable with robust exercise
modalities in clinical trials [1, 6]. More research is required to determine the extent of which
gait dynamics are modifiable to support the creation of strategies to improve gait dynamics
and reduce falls in older adults with MCI and other conditions associated with gait variability
and dynamics.
4.2 IMPLICATIONS AND FUTURE DIRECTIONS
As stated, additional longitudinal research is required to determine the most important and
modifiable characteristics associated with gait dynamics under dual-task conditions, as we
cannot attribute causality to the observations we made in our MCI cohort. Dual-task costs
with regard to gait dynamics have been well characterized, however, which physiological and
neuropsychological attributes contribute independently and most consistently to increased
dual-task cost and the future risk of impaired gait dynamics have not been fully defined. If
modifiable characteristics can be determined, interventions could be developed to target these
specific characteristics to improve gait performance and reduce risk of falls. Current research
is being conducted to determine the link between future falls and worse gait dynamics to
understand what information is needed for clinical assessment in the reduction of falls [7].
Our results suggest that not only physiological, but psychosocial aspects of aging need to be
considered in falls risk assessments. A greater dual-task cost of gait dynamics has been
-132-
associated with lower psychosocial well-being and increased fear of falling [8]. The known
links between fear of falling and physical and mental limitations, reduced social functioning
[9], activity restriction, increased anxiety and depression and reduced physical activity levels
[10] promote the need for clinicians to assess falls risk using a multifactorial assessment tool.
Additionally, associations between depression and prevalence of dementia have been shown
[11], which adds further complexity to the relationship between aging, cognition, gait
dynamics and falling. Thus there is need for clinicians to incorporate a psychosocial
component to the fall risk battery of assessments to determine current well-being, fear of
falling and mood.
In the available literature, there is a lack of consensus regarding the methods used to analyze
gait dynamics [12], which prevents the development of comprehensive recommendations for
dual-task gait testing procedures for use in clinical practice [7]. The development and
utilization of consensus guidelines on gait analysis protocols could identify and stratify fall
risk in cognitively impaired older adults as well as other cohorts at risk [7, 13]. Additionally,
standardizing research methodologies would improve the overall understanding of dual-task
gait performance changes [14] and assist in the identification of risk profiled for clinically
relevant gait dynamics.
A recent review has recommended implementation of combined interventions in order to
improve overall gait for older adults with MCI or early dementia [15]. The term ‘combined
intervention’ refers to the inclusion of strength, balance, and functional mobility training,
along with executive function training [15]. This is concordant with the findings of our acute
dual-task study (Chapter 3) which highlighted specific characteristics that were associated
with dual-task gait performance and have also shown to be modifiable with robust exercise
-133-
modalities in clinical trials. To adequately determine the modifiable and non-modifiable
characteristics of dual-task gait dynamics, more well-designed longitudinal studies and
controlled trials with adequately powered samples are needed, and the independent predictive
value of the clinical characteristics we identified should be explored in multivariable
regression or structural equation modelling. Additionally, more longitudinal research is
required to determine the ideal prescriptions (including exercise and/or other interventions)
needed to promote optimization of gait patterns under single- and dual-task conditions to
enhance and preserve mobility and functional independence in older adults, while minimizing
the risk of falls and related adverse events.
-134-
4.3 REFERENCES
1. Baker, L.D., et al., Effects of Aerobic Exercise on Mild Cognitive Impairment: A
Controlled Trial. Archives of Neurology, 2010. 67(1): p. 71-79.
2. Beauchet, O., et al., Relationship between dual-task related gait changes and intrinsic
risk factors for falls among transitional frail older adults. Aging Clinical and
Experimental Research, 2005. 17(4): p. 270-275.
3. Doi, T., et al., Gray matter volume and dual-task gait performance in mild cognitive
impairment. Brain Imaging and Behavior, 2017. 11(3): p. 887-898.
4. Hausdorff, J.M., et al., Etiology and modification of gait instability in older adults: a
randomized controlled trial of exercise. Journal of Applied Physiology, 2001. 90(6):
p. 2117-2129.
5. Hausdorff, J.M., et al., Dual-Task Decrements in Gait: Contributing Factors Among
Healthy Older Adults. Journals of Gerontology Series a-Biological Sciences and
Medical Sciences, 2008. 63(12): p. 1335-1343.
6. Mavros, Y., et al., Mediation of Cognitive Function Improvements by Strength Gains
After Resistance Training in Older Adults with Mild Cognitive Impairment: Outcomes
of the Study of Mental and Resistance Training. Journal of the American Geriatrics
Society, 2017. 65(3): p. 550-559.
7. Muir-Hunter, S.W. and J.E. Wittwer, Dual-task testing to predict falls in community-
dwelling older adults: a systematic review. Physiotherapy, 2016. 102(1): p. 29-40.
8. Herman, T., et al., Gait instability and fractal dynamics of older adults with a
“cautious” gait: why do certain older adults walk fearfully? Gait & Posture, 2005.
21(2): p. 178-185.
9. Meulen, E., et al., Effect of Fall‐Related Concerns on Physical, Mental, and Social
Function in Community‐Dwelling Older Adults: A Prospective Cohort Study. Journal
of the American Geriatrics Society, 2014. 62(12): p. 2333-2338.
10. Painter, J.A., et al., Fear of Falling and Its Relationship With Anxiety, Depression,
and Activity Engagement Among Community-Dwelling Older Adults. American
Journal of Occupational Therapy, 2012. 66(2): p. 169-176.
-135-
11. Bennett, S. and A.J. Thomas, Depression and dementia: Cause, consequence or
coincidence? Maturitas, 2014. 79(2): p. 184-190.
12. Lord, S., et al., Gait variability in older adults: A structured review of testing protocol
and clinimetric properties. Gait & Posture, 2011. 34(4): p. 443-450.
13. Modarresi, S., et al., Gait parameters and characteristics associated with increased
risk of falls in people with dementia: a systematic review. International
Psychogeriatrics, 2018: p. 1-17.
14. Al-Yahya, E., et al., Cognitive motor interference while walking: A systematic review
and meta-analysis. Neuroscience and Biobehavioral Reviews, 2011. 35(3): p. 715-
728.
15. Zhang, W., et al., Interventions to Improve Gait in Older Adults with Cognitive
Impairment: A Systematic Review. Journal of the American Geriatrics Society, 2019.
67(2): p. 381-391.
-136-
BIBLIOGRAPHY
Al-Yahya, E., Dawes, H., Smith, L., Dennis, A., Howells, K., & Cockburn, J. (2011).
Cognitive motor interference while walking: A systematic review and meta-analysis.
Neuroscience and Biobehavioral Reviews, 35(3), 715-728.
Alberti, K. G., Zimmet, P. Z., & Shaw, J. M. (2006). Metabolic syndrome-- a new world-
wide definition. A Consensus Statement from the International Diabetes Federation.
Diabetic medicine : A journal of the British Diabetic Association, 23(5), 469-480.
Allali, G., Annweiler, C., Blumen, H. M., Callisaya, M. L., De Cock, A. M., Kressig, R. W.,
. . . Beauchet, O. (2016). Gait phenotype from mild cognitive impairment to moderate
dementia: results from the GOOD initiative. European Journal of Neurology, 23(3), 527-
541.
Allali, G., Assal, F., Kressig, R. W., Dubost, V., Herrmann, F. R., & Beauchet, O. (2008).
Impact of impaired executive function on gait stability. Dementia & Geriatric Cognitive
Disorders, 26(4), 364-369.
Allali, G., Dubois, B., Assal, F., Lallart, E., de Souza, L. C., Bertoux, M., . . . Beauchet, O.
(2010). Frontotemporal dementia: pathology of gait? Movement Disorders, 25(6), 731-
737.
Allali, G., Kressig, R. W., Assal, F., Herrmann, F. R., Dubost, V., & Beauchet, O. (2007).
Changes in gait while backward counting in demented older adults with frontal lobe
dysfunction. Gait & Posture, 26(4), 572-576.
Allali, G., Launay, C. P., Blumen, H. M., Callisaya, M. L., De Cock, A.-M., Kressig, R. W.,
. . . Biomathics, C. (2017). Falls, Cognitive Impairment, and Gait Performance: Results
From the GOOD Initiative. Journal of the American Medical Directors Association,
18(4), 335-340.
-137-
Andrews, F. M., & Withey, S. B. (1976). Social indicators of well-being: Americans'
perceptions of life quality. New York: Plenum Press.
Arif, M., Ohtaki, Y., Nagatomi, R., & Inooka, H. (2004). Estimation of the effect of cadence
on gait stability in young and elderly people using approximate entropy technique.
Measurement Science Review, 4(2), 29-40.
Auvinet, B., Touzard, C., Montestruc, F., Delafond, A., & Goeb, V. (2017). Gait disorders in
the elderly and dual task gait analysis: a new approach for identifying motor phenotypes.
Journal of Neuroengineering & Rehabilitation, 14(1), 7.
Ayers, E. I., Tow, A. C., Holtzer, R., & Verghese, J. (2014). Walking while Talking and Falls
in Aging. Gerontology, 60(2), 108-113.
Bahureksa, L., Najafi, B., Saleh, A., Sabbagh, M., Coon, D., Mohler, M. J., & Schwenk, M.
(2017). The Impact of Mild Cognitive Impairment on Gait and Balance: A Systematic
Review and Meta-Analysis of Studies Using Instrumented Assessment. Gerontology,
63(1), 67-83.
Baker, L. D., Frank, L. L., Foster-Schubert, K., Green, P. S., Wilkinson, C. W., McTiernan,
A., . . . Craft, S. (2010). Effects of Aerobic Exercise on Mild Cognitive Impairment: A
Controlled Trial. Archives of Neurology, 67(1), 71-79.
Beauchet, O., Annweiler, C., Callisaya, M. L., De Cock, A.-M., Helbostad, J. L., Kressig, R.
W., . . . Allali, G. (2016). Poor Gait Performance and Prediction of Dementia: Results
From a Meta-Analysis. Journal of the American Medical Directors Association, 17(6),
482-490.
Beauchet, O., Annweiler, C., Dubost, V., Allali, G., Kressig, R. W., Bridenbaugh, S., . . .
Herrmann, F. R. (2009). Stops walking when talking: A predictor of falls in older adults?
European Journal of Neurology, 16(7), 786-795.
-138-
Beauchet, O., Dubost, V., Herrmann, F., Rabilloud, M., Gonthier, R., & Kressig, R. W.
(2005). Relationship between dual-task related gait changes and intrinsic risk factors for
falls among transitional frail older adults. Aging Clinical and Experimental Research,
17(4), 270-275.
Beauchet, O., Freiberger, E., Annweiler, C., Kressig, R., Herrmann, F., & Allali, G. (2011).
Test-retest reliability of stride time variability while dual tasking in healthy and
demented adults with frontotemporal degeneration. Journal of NeuroEngineering and
Rehabilitation, 8(1).
Beauchet, O., Launay, C., Allali, G., Herrmann, F., & Annweiler, C. (2014). Gait Changes
with Anti-Dementia Drugs: A Prospective, Open-Label Study Combining Single and
Dual Task Assessments in Patients with Alzheimer's Disease. Drugs & Aging, 31(5),
363-372.
Beauchet, O., Launay, C. P., Annweiler, C., & Allali, G. (2015). Hippocampal volume, early
cognitive decline and gait variability: Which association? Experimental Gerontology, 61,
98-104.
Beauchet, O., Launay, C. P., Sekhon, H., Barthelemy, J.-C., Roche, F., Chabot, J., . . . Allali,
G. (2017). Association of increased gait variability while dual tasking and cognitive
decline: results from a prospective longitudinal cohort pilot study. GeroScience, 39(4),
439-445.
Bennett, S., & Thomas, A. J. (2014). Depression and dementia: Cause, consequence or
coincidence? Maturitas, 79(2), 184-190.
Bjelland, I., Dahl, A. A., Haug, T. T., & Neckelmann, D. (2002). The validity of the Hospital
Anxiety and Depression Scale: an updated literature review. Journal of psychosomatic
research, 52(2), 69-77.
-139-
Blake, A., Morgan, K., Bendall, M., Dallosso, H., Ebrahim, S., Arie, T., . . . Bassey, E. (1988).
Falls by elderly people at home: prevalence and associated factors. Age and ageing,
17(6), 365-372.
Bloem, B. R., Grimbergen, Y. A. M., van Dijk, J. G., & Munneke, M. (2006). The "posture
second" strategy: A review of wrong priorities in Parkinson's disease. Journal of the
Neurological Sciences, 248(1-2), 196-204.
Booth, V., Hood, V., & Kearney, F. (2016). Interventions incorporating physical and
cognitive elements to reduce falls risk in cognitively impaired older adults: A systematic
review. JBI Database of Systematic Reviews and Implementation Reports, 14(5), 110-
135.
Brach, J. S., Berlin, J. E., VanSwearingen, J. M., Newman, A. B., & Studenski, S. A. (2005).
Too much or too little step width variability is associated with a fall history in older
persons who walk at or near normal gait speed. Journal of NeuroEngineering and
Rehabilitation, 2, 21-21.
Brach, J. S., Studenski, S. A., Perera, S., VanSwearingen, J. M., & Newman, A. B. (2007).
Gait variability and the risk of incident mobility disability in community-dwelling older
adults. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences,
62(9), 983-988.
Bradley, C., & Harrison, J. (2007). Hospitalisations due to falls in older people, Australia,
2003-2004 Injury research and statistics series no. 32 (Vol. Cat. no. INJCAT 96).
Canberra: AIHW.
Burckhardt, C. S., & Anderson, K. L. (2003). The Quality of Life Scale (QOLS): Reliability,
validity, and utilization. Health and Quality of Life Outcomes, 1(1), 60-60.
-140-
Byun, S., Han, J. W., Kim, T. H., Kim, K., Kim, T. H., Park, J. Y., . . . Kim, K. W. (2018).
Gait Variability Can Predict the Risk of Cognitive Decline in Cognitively Normal Older
People. Dementia and Geriatric Cognitive Disorders, 251-261.
Callisaya, M. L., Blizzard, L., Schmidt, M. D., Martin, K. L., McGinley, J. L., Sanders, L.
M., & Srikanth, V. K. (2011). Gait, gait variability and the risk of multiple incident falls
in older people: a population-based study. Age and ageing, 40(4), 481-487.
Camicioli, R., Bouchard, T., & Licis, L. (2006). Dual-tasks and walking fast: relationship to
extra-pyramidal signs in advanced Alzheimer disease. Journal of the Neurological
Sciences, 248(1-2), 205-209.
Camicioli, R., & Licis, L. (2004). Motor impairment predicts falls in specialized Alzheimer
care units. Alzheimer Disease & Associated Disorders, 18(4), 214-218.
Clare, L., Wilson, B. A., Carter, G., Roth, I., & Hodges, J. R. (2002). Assessing awareness in
early-stage Alzheimer's disease: Development and piloting of the Memory Awareness
Rating Scale. Neuropsychological Rehabilitation, 12(4), 341-362.
Clegg, A., Young, J., Iliffe, S., Rikkert, M. O., & Rockwood, K. (2013). Frailty in elderly
people. The Lancet, 381(9868), 752-762.
Cohen, J. (1988). Statistical power analysis for the behavioural sciences (Second edition ed.).
Hillside, NJ, USA: Lawrence Erlbaum Associates.
Corp, D. T., Youssef, G. J., Clark, R. A., Gomes-Osman, J., Yücel, M. A., Oldham, S. J., . .
. Rogers, M. A. (2018). Reduced motor cortex inhibition and a ‘cognitive-first’
prioritisation strategy for older adults during dual-tasking. Experimental Gerontology,
113, 95-105.
D'Esposito, M., Postle, B. R., Ballard, D., & Lease, J. (1999). Maintenance versus
Manipulation of Information Held in Working Memory: An Event-Related fMRI Study.
Brain and Cognition, 41(1), 66-86.
-141-
Delbaere, K., Kochan, N. A., Close, J. C. T., Menant, J. C., Sturnieks, D. L., Brodaty, H., . .
. Lord, S. R. (2012). Mild Cognitive Impairment as a Predictor of Falls in Community-
Dwelling Older People. The American Journal of Geriatric Psychiatry, 20(10), 845-853.
Doi, T., Blumen, H. M., Verghese, J., Shimada, H., Makizako, H., Tsutsumimoto, K., . . .
Suzuki, T. (2017). Gray matter volume and dual-task gait performance in mild cognitive
impairment. Brain Imaging and Behavior, 11(3), 887-898.
Downs, S. H., & Black, N. (1998). The feasibility of creating a checklist for the assessment
of the methodological quality both of randomised and non-randomised studies of health
care interventions. Journal of Epidemiology and Community Health, 52(6), 377-384.
Fiatarone Singh, M. A., Gates, N., Saigal, N., Wilson, G. C., Meiklejohn, J., Brodaty, H., . .
. Valenzuela, M. (2014). The Study of Mental and Resistance Training (SMART)
study—resistance training and/or cognitive training in mild cognitive impairment: a
randomized, double-blind, double-sham controlled trial. Journal of the American
Medical Directors Association, 15(12), 873-880.
Florence, C. S., Bergen, G., Atherly, A., Burns, E., Stevens, J., & Drake, C. (2018). Medical
Costs of Fatal and Nonfatal Falls in Older Adults. Journal of the American Geriatrics
Society, 66(4), 693-698.
Gates, N., Singh, M. A. F., Sachdev, P. S., & Valenzuela, M. (2013). The effect of exercise
training on cognitive function in older adults with mild cognitive impairment: a meta-
analysis of randomized controlled trials. The American Journal of Geriatric Psychiatry,
21(11), 1086-1097.
Gates, N., Valenzuela, M., Sachdev, P. S., & Fiatarone Singh, M. A. (2014). Psychological
well-being in individuals with mild cognitive impairment. Clinical Interventions in
Aging, 9, 779-792.
-142-
Gates, N. J., Valenzuela, M., Sachdev, P. S., Singh, N. A., Baune, B. T., Brodaty, H., . . .
Fiatarone Singh, M. A. (2011). Study of Mental Activity and Regular Training (SMART)
in at risk individuals: A randomised double blind, sham controlled, longitudinal trial.
BMC Geriatrics, 11(1), 19.
Gillain, S., Drame, M., Lekeu, F., Wojtasik, V., Ricour, C., Croisier, J. L., . . . Petermans, J.
(2016). Gait speed or gait variability, which one to use as a marker of risk to develop
Alzheimer disease? A pilot study. Aging-Clinical & Experimental Research, 28(2), 249-
255.
Gillain, S., Warzee, E., Lekeu, F., Wojtasik, V., Maquet, D., Croisier, J. L., . . . Petermans, J.
(2009). The value of instrumental gait analysis in elderly healthy, MCI or Alzheimer's
disease subjects and a comparison with other clinical tests used in single and dual-task
conditions. Annals of Physical & Rehabilitation Medicine, 52(6), 453-474.
Gschwind, Y., Bridenbaugh, S., Reinhard, S., Granacher, U., Monsch, A., & Kressig, R.
(2017). Ginkgo biloba special extract LI 1370 improves dual-task walking in patients
with MCI: a randomised, double-blind, placebo-controlled exploratory study. Aging
clinical and experimental research. Issue, 4, 609-619.
Guralnik, J. M., Simonsick, E. M., Ferrucci, L., Glynn, R. J., Berkman, L. F., Blazer, D. G.,
. . . Wallace, R. B. (1994). A short physical performance battery assessing lower
extremity function: Association with self-reported disability and prediction of mortality
and nursing home admission. Journals of Gerontology, 49(2), M85-M94.
Hamacher, D., Herold, F., Wiegel, P., Hamacher, D., & Schega, L. (2015). Brain activity
during walking: A systematic review. Neuroscience and Biobehavioral Reviews, 57,
310-327.
-143-
Hamacher, D., Singh, N. B., Van Dieen, J. H., Heller, M. O., & Taylor, W. R. (2011).
Kinematic measures for assessing gait stability in elderly individuals: a systematic
review. Journal of the Royal Society Interface, 8(65), 1682-1698.
Hamilton, F., Rochester, L., Paul, L., Rafferty, D., O'Leary, C., & Evans, J. (2009). Walking
and talking: an investigation of cognitive—motor dual tasking in multiple sclerosis.
Multiple Sclerosis Journal, 15(10), 1215-1227.
Hausdorff, J. M. (2005). Gait variability: methods, modeling and meaning. Journal of
NeuroEngineering and Rehabilitation, 2(1), 19.
Hausdorff, J. M. (2007). Gait dynamics, fractals and falls: Finding meaning in the stride-to-
stride fluctuations of human walking. Human Movement Science, 26(4), 555-589.
Hausdorff, J. M., Ashkenazy, Y., Peng, C.-K., Ivanov, P. C., Stanley, H. E., & Goldberger,
A. L. (2001). When human walking becomes random walking: fractal analysis and
modeling of gait rhythm fluctuations. Physica A: Statistical Mechanics and its
Applications, 302(1), 138-147.
Hausdorff, J. M., Balash, J., & Giladi, N. (2003). Effects of Cognitive Challenge on Gait
Variability in Patients with Parkinson’s Disease. Journal of Geriatric Psychiatry and
Neurology, 16(1), 53-58.
Hausdorff, J. M., Edelberg, H. K., Mitchell, S. L., Goldberger, A. L., & Wei, J. Y. (1997).
Increased gait unsteadiness in community-dwelling elderly fallers. Archives of Physical
Medicine and Rehabilitation, 78(3), 278-283.
Hausdorff, J. M., Ladin, Z., & Wei, J. Y. (1995). Footswitch system for measurement of the
temporal parameters of gait. Journal of Biomechanics, 28(3), 347-351.
Hausdorff, J. M., Nelson, M. E., Kaliton, D., Layne, J. E., Bernstein, M. J., Nuernberger, A.,
& Fiatarone Singh, M. A. (2001). Etiology and modification of gait instability in older
-144-
adults: a randomized controlled trial of exercise. Journal of Applied Physiology, 90(6),
2117-2129.
Hausdorff, J. M., Peng, C. K., Ladin, Z., Wei, J. Y., & Goldberger, A. L. (1995). Is walking
a random walk? Evidence for long-range correlations in stride interval of human gait.
Journal of Applied Physiology, 78(1), 349-358.
Hausdorff, J. M., Purdon, P. L., Peng, C. K., Ladin, Z., Wei, J. Y., & Goldberger, A. L.
(1996). Fractal dynamics of human gait: stability of long-range correlations in stride
interval fluctuations. Journal of Applied Physiology, 80(5), 1448-1457.
Hausdorff, J. M., Rios, D. A., & Edelberg, H. K. (2001). Gait variability and fall risk in
community-living older adults: A 1-year prospective study. Archives of Physical
Medicine and Rehabilitation, 82(8), 1050-1056.
Hausdorff, J. M., Schweiger, A., Herman, T., Yogev-Seligmann, G., & Giladi, N. (2008).
Dual-Task Decrements in Gait: Contributing Factors Among Healthy Older Adults.
Journals of Gerontology Series a-Biological Sciences and Medical Sciences, 63(12),
1335-1343.
Hausdorff, J. M., Yogev, G., Springer, S., Simon, E. S., & Giladi, N. (2005). Walking is more
like catching than tapping: gait in the elderly as a complex cognitive task. Experimental
Brain Research, 164(4), 541-548.
Hedges, L. V. (1981). Distribution theory for Glass's estimator of effect size and related
estimators. journal of Educational Statistics, 6(2), 107-128.
Herman, T., Giladi, N., Gurevich, T., & Hausdorff, J. M. (2005). Gait instability and fractal
dynamics of older adults with a “cautious” gait: why do certain older adults walk
fearfully? Gait & Posture, 21(2), 178-185.
-145-
Herssens, N., Verbecque, E., Hallemans, A., Vereeck, L., Van Rompaey, V., & Saeys, W.
(2018). Do spatiotemporal parameters and gait variability differ across the lifespan of
healthy adults? A systematic review. Gait & Posture, 64, 181-190.
Higgins, J., & Green, S. (editors). (2011). Cochrane Handbook for Systematic Reviews of
Interventions, Version 5.1.0 [updated March 2011]. The Cochrane Collaboration.
Retrieved from Available from www.handbook.cochrane.org.
Hsu, Y. L., Chung, P. C., Wang, W. H., Pai, M. C., Wang, C. Y., Lin, C. W., . . . Wang, J. S.
(2014). Gait and balance analysis for patients with Alzheimer's disease using an inertial-
sensor-based wearable instrument. Journal of Biomedical & Health Informatics, 18(6),
1822-1830.
Ijmker, T., & Lamoth, C. J. (2012). Gait and cognition: the relationship between gait stability
and variability with executive function in persons with and without dementia. Gait &
Posture, 35(1), 126-130.
Jansen, I., Heymsfield, S., Baumgartner, R., & Ross, R. (2000). Estimation of skeletal muscle
mass by bioelectrical impedance analysis. Journal of Applied Physiology, 89(2), 465-
471.
Kaufman, A. S., & Lichtenberger, E. O. (1999). Essentials: of WAIS-III assessment. New
York: J. Wiley & Sons.
Kelly, V. E., Eusterbrock, A. J., & Shumway-Cook, A. (2012). A Review of Dual-Task
Walking Deficits in People with Parkinson’s Disease: Motor and Cognitive
Contributions, Mechanisms, and Clinical Implications. Parkinson's Disease, 2012(1), 14.
Kikkert, L., Vuillerme, N., van Campen, J. P., Appels, B. A., Hortobagyi, T., & Lamoth, C.
J. (2017). Gait characteristics and their discriminative power in geriatric patients with
and without cognitive impairment. Journal of NeuroEngineering and Rehabilitation,
14(1), 84.
-146-
Koenig, H. G., Westlund, R. E., George, L. K., Hughes, D. C., Blazer, D. G., & Hybels, C.
(1993). Abbreviating the Duke Social Support Index for use in chronically ill elderly
individuals. Psychosomatics, 34(1), 61-69.
König, A., Klaming, L., Pijl, M., Demeurraux, A., David, R., & Robert, P. (2017). Objective
measurement of gait parameters in healthy and cognitively impaired elderly using the
dual-task paradigm. Aging Clinical & Experimental Research, 29(6), 1181-1189.
Kvelde, T., Lord, S. R., Close, J. C. T., Reppermund, S., Kochan, N. A., Sachdev, P., . . .
Delbaere, K. (2015). Depressive symptoms increase fall risk in older people, independent
of antidepressant use, and reduced executive and physical functioning. Archives of
Gerontology and Geriatrics, 60(1), 190-195.
Lamoth, C. J., van Deudekom, F. J., van Campen, J. P., Appels, B. A., de Vries, O. J., &
Pijnappels, M. (2011). Gait stability and variability measures show effects of impaired
cognition and dual tasking in frail people. Journal of Neuroengineering & Rehabilitation,
8, 2.
Leal, S. L., & Yassa, M. A. (2013). Perturbations of neural circuitry in aging, mild cognitive
impairment, and Alzheimer's disease. Ageing Research Reviews, 12(3), 823-831.
Lee, J., & Park, S. (2018). Effects of a priority-based dual task on gait velocity and variability
in older adults with mild cognitive impairment. Journal of Exercise Rehabilitation, 14(6),
993-997.
Leech, R., Kamourieh, S., Beckmann, C. F., & Sharp, D. J. (2011). Fractionating the Default
Mode Network: Distinct Contributions of the Ventral and Dorsal Posterior Cingulate
Cortex to Cognitive Control. The Journal of Neuroscience, 31(9), 3217-3224.
Leech, R., & Sharp, D. J. (2014). The role of the posterior cingulate cortex in cognition and
disease. Brain, 137(1), 12-32.
-147-
Lehrner, J., Gufler, R., Guttmann, G., Maly, J., Gleiß, A., Auff, E., & Dal-Bianco, P. (2005).
Annual conversion to Alzheimer disease among patients with memory complaints
attending an outpatient memory clinic: the influence of amnestic mild cognitive
impairment and the predictive value of neuropsychological testing. Wiener Klinische
Wochenschrift, 117(18), 629-635.
Lezak, M. D., Howieson, D. B., Loring, D. W., & Fischer, J. S. (2004). Neuropsychological
assessment: Oxford University Press, USA.
Li, K. Z. H., Bherer, L., Mirelman, A., Maidan, I., & Hausdorff, J. M. (2018). Cognitive
Involvement in Balance, Gait and Dual-Tasking in Aging: A Focused Review From a
Neuroscience of Aging Perspective. Frontiers in Neurology, 9(1), 913.
Lin, Y. C., Hsu, W. C., Wu, C. K., Chang, W. H., Wu, K. P. H., & Wong, A. M. K. (2016).
Comparison of motor performance of upper and lower extremities in dual-task tests in
patients with mild Alzheimer's dementia. Aging Clinical and Experimental Research,
28(3), 491-496.
Liu-Ambrose, T. Y., Ashe, M. C., Graf, P., Beattie, B. L., & Khan, K. M. (2008). Increased
risk of falling in older community-dwelling women with mild cognitive impairment.
Physical therapy, 88(12), 1482-1491.
Lord, S., Howe, T., Greenland, J., Simpson, L., & Rochester, L. (2011). Gait variability in
older adults: A structured review of testing protocol and clinimetric properties. Gait &
Posture, 34(4), 443-450.
Lukaski, H., Bolonchuk, W., Hall, C., & Siders, W. (1986). Validation of tetrapolar
bioelectrical impedance method to assess human body composition. Journal of Applied
Physiology, 60(4), 1327-1332.
Lundin-Olsson, L., Nyberg, L., & Gustafson, Y. (1997). “Stops walking when talking” as a
predictor of falls in elderly people. The Lancet, 349(9052), 617.
-148-
Maki, B. E. (1997). Gait Changes in Older Adults: Predictors of Falls or Indicators of Fear?
Journal of the American Geriatrics Society, 45(3), 313-320.
Maquet, D., Lekeu, F., Warzee, E., Gillain, S., Wojtasik, V., Salmon, E., . . . Croisier, J. L.
(2010). Gait analysis in elderly adult patients with mild cognitive impairment and
patients with mild Alzheimer's disease: simple versus dual task: a preliminary report.
Clinical Physiology & Functional Imaging, 30(1), 51-56.
Martin, K. L., Blizzard, L., Wood, A. G., Srikanth, V., Thomson, R., Sanders, L. M., &
Callisaya, M. L. (2013). Cognitive Function, Gait, and Gait Variability in Older People:
A Population-Based Study. The Journals of Gerontology: Series A, 68(6), 726-732.
Martinez-Ramirez, A., Martinikorena, I., Lecumberri, P., Gomez, M., Millor, N., Casas-
Herrero, A., . . . Izquierdo, M. (2016). Dual Task Gait Performance in Frail Individuals
with and without Mild Cognitive Impairment. Dementia & Geriatric Cognitive
Disorders, 42(1-2), 7-16.
Mavros, Y., Gates, N., Wilson, G. C., Jain, N., Meiklejohn, J., Brodaty, H., . . . Fiatarone
Singh, M. A. (2017). Mediation of Cognitive Function Improvements by Strength Gains
After Resistance Training in Older Adults with Mild Cognitive Impairment: Outcomes
of the Study of Mental and Resistance Training. Journal of the American Geriatrics
Society, 65(3), 550-559.
McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., & Stadlan, E. M. (1984).
Clinical diagnosis of Alzheimer's disease Report of the NINCDS ‐ADRDA W o
Group* under the auspices of Department of Health and Human Services Task Force on
Alzheimer's Disease. Neurology, 34(7), 939-939.
Menant, J. C., Schoene, D., Sarofim, M., & Lord, S. R. (2014). Single and dual task tests of
gait speed are equivalent in the prediction of falls in older people: a systematic review
and meta-analysis. Ageing Research Reviews, 16, 83-104.
-149-
Meulen, E., Zijlstra, G., Ambergen, T., & Kempen, G. I. (2014). Effect of Fall ‐Related
Concerns on Physical, Mental, and Social Function in Community ‐Dwelling Ol
Adults: A Prospective Cohort Study. Journal of the American Geriatrics Society, 62(12),
2333-2338.
Meyer, J. S., Xu, G., Thornby, J., Chowdhury, M. H., & Quach, M. (2002). Is Mild Cognitive
Impairment Prodromal for Vascular Dementia Like Alzheimer's Disease? Stroke, 33(8),
1981-1985.
Modarresi, S., Divine, A., Grahn, J. A., Overend, T. J., & Hunter, S. W. (2018). Gait
parameters and characteristics associated with increased risk of falls in people with
dementia: a systematic review. International Psychogeriatrics, 1-17.
Montero-Odasso, M., Bergman, H., Phillips, N. A., Wong, C. H., Sourial, N., & Chertkow,
H. (2009). Dual-tasking and gait in people with mild cognitive impairment. The effect of
working memory. BMC Geriatrics, 9(1), 41.
Montero-Odasso, M., Casas, A., Hansen, K. T., Bilski, P., Gutmanis, I., Wells, J. L., & Borrie,
M. J. (2009). Quantitative gait analysis under dual-task in older people with mild
cognitive impairment: a reliability study. Journal of Neuroengineering & Rehabilitation,
6(1), 35.
Montero-Odasso, M., Muir, S. W., & Speechley, M. (2012). Dual-task complexity affects
gait in people with mild cognitive impairment: the interplay between gait variability, dual
tasking, and risk of falls. Archives of Physical Medicine & Rehabilitation, 93(2), 293-
299.
Montero-Odasso, M., Oteng-Amoako, A., Speechley, M., Gopaul, K., Beauchet, O.,
Annweiler, C., & Muir-Hunter, S. W. (2014). The motor signature of mild cognitive
impairment: results from the gait and brain study. Journals of Gerontology Series A-
Biological Sciences & Medical Sciences, 69(11), 1415-1421.
-150-
Montero-Odasso, M., Verghese, J., Beauchet, O., & Hausdorff, J. M. (2012). Gait and
cognition: A complementary approach to understanding brain function and the risk of
falling. Journal of the American Geriatrics Society, 60(11), 2127-2136.
Moon, Y., Sung, J., An, R., Hernandez, M. E., & Sosnoff, J. J. (2016). Gait variability in
people with neurological disorders: A systematic review and meta-analysis. Human
Movement Science, 47, 197-208.
Muir-Hunter, S. W., & Wittwer, J. E. (2016). Dual-task testing to predict falls in community-
dwelling older adults: a systematic review. Physiotherapy, 102(1), 29-40.
Muir, S. W., Gopaul, K., & Montero Odasso, M. M. (2012). The role of cognitive impairment
in fall risk among older adults: a systematic review and meta-analysis. Age and ageing,
41(3), 299-308.
Muir, S. W., Speechley, M., Wells, J., Borrie, M., Gopaul, K., & Montero-Odasso, M. (2012).
Gait assessment in mild cognitive impairment and Alzheimer's disease: The effect of
dual-task challenges across the cognitive spectrum. Gait and Posture, 35(1), 96-100.
Nakagawa, S. (2004). A farewell to Bonferroni: the problems of low statistical power and
publication bias. Behavioral Ecology, 15(6), 1044-1045.
Nascimbeni, A., Caruso, S., Salatino, A., Carenza, M., Rigano, M., Raviolo, A., & Ricci, R.
(2015). Dual task-related gait changes in patients with mild cognitive impairment.
Functional Neurology, 30(1), 59-65.
Nasreddine, Z. S., Phillips, N. A., Bédirian, V., Charbonneau, S., Whitehead, V., Collin, I., .
. . Chertkow, H. (2005). The Montreal Cognitive Assessment, MoCA: A Brief Screening
Tool For Mild Cognitive Impairment. Journal of the American Geriatrics Society, 53(4),
695-699.
-151-
Ng, F., Trauer, T., Dodd, S., Callaly, T., Campbell, S., & Berk, M. (2007). The validity of
the 21-item version of the Depression Anxiety Stress Scales as a routine clinical outcome
measure. Acta Neuropsychiatrica, 19(5), 304-310.
Painter, J. A., Allison, L., Dhingra, P., Daughtery, J., Cogdill, K., & Trujillo, L. G. (2012).
Fear of Falling and Its Relationship With Anxiety, Depression, and Activity Engagement
Among Community-Dwelling Older Adults. American Journal of Occupational Therapy,
66(2), 169-176.
Petersen, R. C. (1999). Mild Cognitive Impairment: Clinical Characterization and Outcome.
The Journal of the American Medical Association, 281(19), 1777.
Petersen, R. C., Doody, R., Kurz, A., & et al. (2001). Current concepts in mild cognitive
impairment. Archives of Neurology, 58(12), 1985-1992.
Pettersson, A., Olsson, E., & Wahlund, L.-O. (2005). Motor function in subjects with mild
cognitive impairment and early Alzheimer’s disease. Dementia and Geriatric Cognitive
Disorders, 19(5-6), 299-304.
Pirker, W., & Katzenschlager, R. (2017). Gait disorders in adults and the elderly. Wiener
klinische Wochenschrift, 129(3), 81-95.
Plummer, P., Zukowski, L. A., Giuliani, C., Hall, A. M., & Zurakowski, D. (2015). Effects
of Physical Exercise Interventions on Gait-Related Dual-Task Interference in Older
Adults: A Systematic Review and Meta-Analysis. Gerontology, 62(1), 94-117.
Pointer, S. C. (2018). Trends in hospitalised injury, Australia 1999-00 to 2014-15 Injury
research and statistics series no. 110 (Vol. Cat. no. INJCAT 190). Canberra: AIHW.
Rabin, L. A., Paré, N., Saykin, A. J., Brown, M. J., Wishart, H. A., Flashman, L. A., &
Santulli, R. B. (2009). Differential memory test sensitivity for diagnosing amnestic mild
cognitive impairment and predicting conversion to Alzheimer's disease.
-152-
Neuropsychology, development, and cognition. Section B, Aging, neuropsychology and
cognition, 16(3), 357-376.
Rhea, C. K., Silver, T. A., Hong, S. L., Ryu, J. H., Studenka, B. E., Hughes, C. M. L., &
Haddad, J. M. (2011). Noise and Complexity in Human Postural Control: Interpreting
the Different Estimations of Entropy. PLoS ONE [Electronic Resource], 6(3), e17696.
Rosano, C., Aizenstein, H., Brach, J., Longenberger, A., Studenski, S., & Newman, A. B.
(2008). Special ArticleGait Measures Indicate Underlying Focal Gray Matter Atrophy in
the Brain of Older Adults. The Journals of Gerontology: Series A, 63(12), 1380-1388.
Rossiter-Fornoff, J. E., Wolf, S. L., Wolfson, L. I., & Buchner, D. M. (1995). A cross-
sectional validation study of the FICSIT common data base static balance measures.
Frailty and Injuries: Cooperative Studies of Intervention Techniques. The journals of
gerontology. Series A, Biological sciences and medical sciences, 50(6), M291.
Royall, D. R., Lauterbach, E. C., Cummings, J. L., Reeve, A., Rummans, T. A., Kaufer, D.
I., . . . Coffey, C. E. (2002). Executive Control Function. The Journal of Neuropsychiatry
and Clinical Neurosciences, 14(4), 377-405.
Rubenstein, L. Z. (2006). Falls in older people: epidemiology, risk factors and strategies for
prevention. Age and ageing, 35(suppl_2), ii37-ii41.
Rucco, R., Agosti, V., Jacini, F., Sorrentino, P., Varriale, P., De Stefano, M., . . . Sorrentino,
G. (2017). Spatio-temporal and kinematic gait analysis in patients with Frontotemporal
dementia and Alzheimer's disease through 3D motion capture. Gait & Posture, 52, 312-
317.
Ryff, C. D., & Keyes, C. L. M. (1995). The Structure of Psychological Well-Being Revisited.
Journal of Personality and Social Psychology, 69(4), 719-727. doi: 10.1037/0022-
3514.69.4.719
-153-
Schwenk, M., Zieschang, T., Englert, S., Grewal, G., Najafi, B., & Hauer, K. (2014).
Improvements in gait characteristics after intensive resistance and functional training in
people with dementia: a randomised controlled trial. BMC Geriatrics, 14(1), 73.
Shaw, F. E. (2002). Falls in cognitive impairment and dementia. Clinics in Geriatric
Medicine, 18(2), 159-173.
Sheline, Y. I., Wang, P. W., Gado, M. H., Csernansky, J. G., & Vannier, M. W. (1996).
Hippocampal atrophy in recurrent major depression. Proceedings of the National
Academy of Sciences, 93(9), 3908-3913.
Sheridan, P. L., Solomont, J., Kowall, N., & Hausdorff, J. M. (2003). Influence of executive
function on locomotor function: divided attention increases gait variability in
Alzheimer's disease. Journal of the American Geriatrics Society, 51(11), 1633-1637.
Smith, E., Cusack, T., Cunningham, C., & Blake, C. (2017). The Influence of a Cognitive
Dual Task on the Gait Parameters of Healthy Older Adults: A Systematic Review and
Meta-Analysis. Journal of Aging & Physical Activity, 25(4), 671-686.
Snijders, A. H., van de Warrenburg, B. P., Giladi, N., & Bloem, B. R. (2007). Neurological
gait disorders in elderly people: clinical approach and classification. The Lancet
Neurology, 6(1), 63-74.
Springer, S., Giladi, N., Peretz, C., Yogev, G., Simon, E. S., & Hausdorff, J. M. (2006). Dual-
tasking effects on gait variability: The role of aging, falls, and executive function.
Movement Disorders, 21(7), 950-957.
St Clair-Thompson, H. L., & Allen, R. J. (2013). Are forward and backward recall the same?
A dual-task study of digit recall. Memory & Cognition, 41(4), 519-532.
Stergiou, N., & Decker, L. M. (2011). Human movement variability, nonlinear dynamics, and
pathology: Is there a connection? Human Movement Science, 30(5), 869-888.
-154-
Stewart Williams, J., Kowal, P., Hestekin, H., O’Driscoll, T., Peltzer, K., Yawson, A., . . .
collaborators, S. (2015). Prevalence, risk factors and disability associated with fall-
related injury in older adults in low- and middle-incomecountries: results from the WHO
Study on global AGEing and adult health (SAGE). BMC Medicine, 13(1), 147.
Suo, C., Fiatarone Singh, M., Gates, N., Wen, W., Sachdev, P., Brodaty, H., . . . Singh, N.
(2016). Therapeutically relevant structural and functional mechanisms triggered by
physical and cognitive exercise. Molecular Psychiatry, 21(11), 1633.
Taylor, M. E., Delbaere, K., Mikolaizak, A. S., Lord, S. R., & Close, J. C. (2013). Gait
parameter risk factors for falls under simple and dual task conditions in cognitively
impaired older people. Gait & Posture, 37(1), 126-130.
Tian, Q., Chastan, N., Bair, W.-N., Resnick, S. M., Ferrucci, L., & Studenski, S. A. (2017).
The brain map of gait variability in aging, cognitive impairment and dementia—a
systematic review. Neuroscience & Biobehavioral Reviews, 74, 149-162.
van Iersel, M. B., Hoefsloot, W., Munneke, M., Bloem, B. R., & Olde Rikkert, M. G. M.
(2004). Systematic review of quantitativeclinical gait analysis in patients with dementia.
Zeitschrift für Gerontologie und Geriatrie, 37(1), 27-32.
Verghese, J., LeValley, A., Hall, C. B., Katz, M. J., Ambrose, A. F., & Lipton, R. B. (2006).
Epidemiology of Gait Disorders in Community-Residing Older Adults. Journal of the
American Geriatrics Society, 54(2), 255-261.
Verhaeghen, P., Steitz, D. W., Silwinski, M. J., & Cerella, J. (2003). Aging and Dual-Task
Performance: A Meta-Analysis. Psychology and Aging, 18(3), 443-460.
Wajda, D. A., Mirelman, A., Hausdorff, J. M., & Sosnoff, J. J. (2017). Intervention modalities
for targeting cognitive-motor interference in individuals with neurodegenerative disease:
a systematic review. Expert Review of Neurotherapeutics, 17(3), 251-261.
-155-
Whooley, M. A., Kip, K. E., Cauley, J. A., Ensrud, K. E., Nevitt, M. C., Browner, W. S., &
Group, f. t. S. o. O. F. R. (1999). Depression, Falls, and Risk of Fracture in Older Women.
Archives of Internal Medicine, 159(5), 484-490.
Winblad, B., Palmer, K., Kivipelto, M., Jelic, V., Fratiglioni, L., Wahlund, L.-O., . . .
Petersen, R. C. (2004). Mild cognitive impairment – beyond controversies, towards a
consensus: report of the International Working Group on Mild Cognitive Impairment.
Journal of Internal Medicine, 256(3), 240-246.
Wittwer, J. E., Webster, K. E., & Hill, K. (2014). The effects of a concurrent motor task on
walking in Alzheimer's disease. Gait & Posture, 39(1), 291-296.
Yesavage, J. A., Brink, T. L., Rose, T. L., Lum, O., Huang, V., Adey, M., & Leirer, V. O.
(1982). Development and validation of a geriatric depression screening scale: a
preliminary report. Journal of psychiatric research, 17(1), 37-49.
Yogev-Seligmann, G., Hausdorff, J. M., & Giladi, N. (2008). The role of executive function
and attention in gait. Movement Disorders, 23(3), 329-342.
Zhang, W., Low, L.-F., Gwynn, J. D., & Clemson, L. (2019). Interventions to Improve Gait
in Older Adults with Cognitive Impairment: A Systematic Review. Journal of the
American Geriatrics Society, 67(2), 381-391.
Zimmerman, M. E., Lipton, R. B., Pan, J. W., Hetherington, H. P., & Verghese, J. (2009).
MRI- and MRS-Derived Hippocampal Correlates of Quantitative Locomotor Function
in Older Adults. Brain research, 1291, 73-81.